mariadb/sql/item_cmpfunc.h

1792 lines
54 KiB
C
Raw Normal View History

#ifndef ITEM_CMPFUNC_INCLUDED
#define ITEM_CMPFUNC_INCLUDED
2012-02-16 10:48:16 +01:00
/* Copyright (c) 2000, 2011, Oracle and/or its affiliates. All rights reserved.
2000-07-31 21:29:14 +02:00
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
2000-07-31 21:29:14 +02:00
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
2000-07-31 21:29:14 +02:00
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
/* compare and test functions */
#ifdef USE_PRAGMA_INTERFACE
2000-07-31 21:29:14 +02:00
#pragma interface /* gcc class implementation */
#endif
#include "thr_malloc.h" /* sql_calloc */
#include "item_func.h" /* Item_int_func, Item_bool_func */
#include "my_regex.h"
extern Item_result item_cmp_type(Item_result a,Item_result b);
class Item_bool_func2;
class Arg_comparator;
typedef int (Arg_comparator::*arg_cmp_func)();
typedef int (*Item_field_cmpfunc)(Item_field *f1, Item_field *f2, void *arg);
class Arg_comparator: public Sql_alloc
{
Item **a, **b;
arg_cmp_func func;
Item_result_field *owner;
Arg_comparator *comparators; // used only for compare_row()
double precision;
Bug#27590: Wrong DATE/DATETIME comparison. DATE and DATETIME can be compared either as strings or as int. Both methods have their disadvantages. Strings can contain valid DATETIME value but have insignificant zeros omitted thus became non-comparable with other DATETIME strings. The comparison as int usually will require conversion from the string representation and the automatic conversion in most cases is carried out in a wrong way thus producing wrong comparison result. Another problem occurs when one tries to compare DATE field with a DATETIME constant. The constant is converted to DATE losing its precision i.e. losing time part. This fix addresses the problems described above by adding a special DATE/DATETIME comparator. The comparator correctly converts DATE/DATETIME string values to int when it's necessary, adds zero time part (00:00:00) to DATE values to compare them correctly to DATETIME values. Due to correct conversion malformed DATETIME string values are correctly compared to other DATE/DATETIME values. As of this patch a DATE value equals to DATETIME value with zero time part. For example '2001-01-01' equals to '2001-01-01 00:00:00'. The compare_datetime() function is added to the Arg_comparator class. It implements the correct comparator for DATE/DATETIME values. Two supplementary functions called get_date_from_str() and get_datetime_value() are added. The first one extracts DATE/DATETIME value from a string and the second one retrieves the correct DATE/DATETIME value from an item. The new Arg_comparator::can_compare_as_dates() function is added and used to check whether two given items can be compared by the compare_datetime() comparator. Two caching variables were added to the Arg_comparator class to speedup the DATE/DATETIME comparison. One more store() method was added to the Item_cache_int class to cache int values. The new is_datetime() function was added to the Item class. It indicates whether the item returns a DATE/DATETIME value.
2007-04-26 22:12:09 +02:00
/* Fields used in DATE/DATETIME comparison. */
THD *thd;
enum_field_types a_type, b_type; // Types of a and b items
Item *a_cache, *b_cache; // Cached values of a and b items
bool is_nulls_eq; // TRUE <=> compare for the EQUAL_FUNC
bool set_null; // TRUE <=> set owner->null_value
// when one of arguments is NULL.
Bug#27590: Wrong DATE/DATETIME comparison. DATE and DATETIME can be compared either as strings or as int. Both methods have their disadvantages. Strings can contain valid DATETIME value but have insignificant zeros omitted thus became non-comparable with other DATETIME strings. The comparison as int usually will require conversion from the string representation and the automatic conversion in most cases is carried out in a wrong way thus producing wrong comparison result. Another problem occurs when one tries to compare DATE field with a DATETIME constant. The constant is converted to DATE losing its precision i.e. losing time part. This fix addresses the problems described above by adding a special DATE/DATETIME comparator. The comparator correctly converts DATE/DATETIME string values to int when it's necessary, adds zero time part (00:00:00) to DATE values to compare them correctly to DATETIME values. Due to correct conversion malformed DATETIME string values are correctly compared to other DATE/DATETIME values. As of this patch a DATE value equals to DATETIME value with zero time part. For example '2001-01-01' equals to '2001-01-01 00:00:00'. The compare_datetime() function is added to the Arg_comparator class. It implements the correct comparator for DATE/DATETIME values. Two supplementary functions called get_date_from_str() and get_datetime_value() are added. The first one extracts DATE/DATETIME value from a string and the second one retrieves the correct DATE/DATETIME value from an item. The new Arg_comparator::can_compare_as_dates() function is added and used to check whether two given items can be compared by the compare_datetime() comparator. Two caching variables were added to the Arg_comparator class to speedup the DATE/DATETIME comparison. One more store() method was added to the Item_cache_int class to cache int values. The new is_datetime() function was added to the Item class. It indicates whether the item returns a DATE/DATETIME value.
2007-04-26 22:12:09 +02:00
enum enum_date_cmp_type { CMP_DATE_DFLT= 0, CMP_DATE_WITH_DATE,
CMP_DATE_WITH_STR, CMP_STR_WITH_DATE };
longlong (*get_value_a_func)(THD *thd, Item ***item_arg, Item **cache_arg,
Item *warn_item, bool *is_null);
longlong (*get_value_b_func)(THD *thd, Item ***item_arg, Item **cache_arg,
Item *warn_item, bool *is_null);
bool try_year_cmp_func(Item_result type);
public:
DTCollation cmp_collation;
/* Allow owner function to use string buffers. */
String value1, value2;
Arg_comparator(): comparators(0), thd(0), a_cache(0), b_cache(0), set_null(TRUE),
get_value_a_func(0), get_value_b_func(0) {};
Arg_comparator(Item **a1, Item **a2): a(a1), b(a2), comparators(0), thd(0),
a_cache(0), b_cache(0), set_null(TRUE),
get_value_a_func(0), get_value_b_func(0) {};
int set_compare_func(Item_result_field *owner, Item_result type);
inline int set_compare_func(Item_result_field *owner_arg)
{
return set_compare_func(owner_arg, item_cmp_type((*a)->result_type(),
2006-06-14 23:48:41 +02:00
(*b)->result_type()));
}
int set_cmp_func(Item_result_field *owner_arg,
Item **a1, Item **a2,
Bug#27590: Wrong DATE/DATETIME comparison. DATE and DATETIME can be compared either as strings or as int. Both methods have their disadvantages. Strings can contain valid DATETIME value but have insignificant zeros omitted thus became non-comparable with other DATETIME strings. The comparison as int usually will require conversion from the string representation and the automatic conversion in most cases is carried out in a wrong way thus producing wrong comparison result. Another problem occurs when one tries to compare DATE field with a DATETIME constant. The constant is converted to DATE losing its precision i.e. losing time part. This fix addresses the problems described above by adding a special DATE/DATETIME comparator. The comparator correctly converts DATE/DATETIME string values to int when it's necessary, adds zero time part (00:00:00) to DATE values to compare them correctly to DATETIME values. Due to correct conversion malformed DATETIME string values are correctly compared to other DATE/DATETIME values. As of this patch a DATE value equals to DATETIME value with zero time part. For example '2001-01-01' equals to '2001-01-01 00:00:00'. The compare_datetime() function is added to the Arg_comparator class. It implements the correct comparator for DATE/DATETIME values. Two supplementary functions called get_date_from_str() and get_datetime_value() are added. The first one extracts DATE/DATETIME value from a string and the second one retrieves the correct DATE/DATETIME value from an item. The new Arg_comparator::can_compare_as_dates() function is added and used to check whether two given items can be compared by the compare_datetime() comparator. Two caching variables were added to the Arg_comparator class to speedup the DATE/DATETIME comparison. One more store() method was added to the Item_cache_int class to cache int values. The new is_datetime() function was added to the Item class. It indicates whether the item returns a DATE/DATETIME value.
2007-04-26 22:12:09 +02:00
Item_result type);
inline int set_cmp_func(Item_result_field *owner_arg,
Item **a1, Item **a2, bool set_null_arg)
{
set_null= set_null_arg;
2006-06-14 23:48:41 +02:00
return set_cmp_func(owner_arg, a1, a2,
item_cmp_type((*a1)->result_type(),
(*a2)->result_type()));
}
inline int compare() { return (this->*func)(); }
int compare_string(); // compare args[0] & args[1]
int compare_binary_string(); // compare args[0] & args[1]
int compare_real(); // compare args[0] & args[1]
2005-02-08 23:50:45 +01:00
int compare_decimal(); // compare args[0] & args[1]
int compare_int_signed(); // compare args[0] & args[1]
int compare_int_signed_unsigned();
int compare_int_unsigned_signed();
int compare_int_unsigned();
int compare_row(); // compare args[0] & args[1]
int compare_e_string(); // compare args[0] & args[1]
int compare_e_binary_string(); // compare args[0] & args[1]
int compare_e_real(); // compare args[0] & args[1]
2005-02-08 23:50:45 +01:00
int compare_e_decimal(); // compare args[0] & args[1]
int compare_e_int(); // compare args[0] & args[1]
int compare_e_int_diff_signedness();
int compare_e_row(); // compare args[0] & args[1]
int compare_real_fixed();
int compare_e_real_fixed();
Bug#27590: Wrong DATE/DATETIME comparison. DATE and DATETIME can be compared either as strings or as int. Both methods have their disadvantages. Strings can contain valid DATETIME value but have insignificant zeros omitted thus became non-comparable with other DATETIME strings. The comparison as int usually will require conversion from the string representation and the automatic conversion in most cases is carried out in a wrong way thus producing wrong comparison result. Another problem occurs when one tries to compare DATE field with a DATETIME constant. The constant is converted to DATE losing its precision i.e. losing time part. This fix addresses the problems described above by adding a special DATE/DATETIME comparator. The comparator correctly converts DATE/DATETIME string values to int when it's necessary, adds zero time part (00:00:00) to DATE values to compare them correctly to DATETIME values. Due to correct conversion malformed DATETIME string values are correctly compared to other DATE/DATETIME values. As of this patch a DATE value equals to DATETIME value with zero time part. For example '2001-01-01' equals to '2001-01-01 00:00:00'. The compare_datetime() function is added to the Arg_comparator class. It implements the correct comparator for DATE/DATETIME values. Two supplementary functions called get_date_from_str() and get_datetime_value() are added. The first one extracts DATE/DATETIME value from a string and the second one retrieves the correct DATE/DATETIME value from an item. The new Arg_comparator::can_compare_as_dates() function is added and used to check whether two given items can be compared by the compare_datetime() comparator. Two caching variables were added to the Arg_comparator class to speedup the DATE/DATETIME comparison. One more store() method was added to the Item_cache_int class to cache int values. The new is_datetime() function was added to the Item class. It indicates whether the item returns a DATE/DATETIME value.
2007-04-26 22:12:09 +02:00
int compare_datetime(); // compare args[0] & args[1] as DATETIMEs
static enum enum_date_cmp_type can_compare_as_dates(Item *a, Item *b,
ulonglong *const_val_arg);
Bug#34384: Slow down on constant conversion. When values of different types are compared they're converted to a type that allows correct comparison. This conversion is done for each comparison and takes some time. When a constant is being compared it's possible to cache the value after conversion to speedup comparison. In some cases (large dataset, complex WHERE condition with many type conversions) query might be executed 7% faster. A test case isn't provided because all changes are internal and isn't visible outside. The behavior of the Item_cache is changed to cache values on the first request of cached value rather than at the moment of storing item to be cached. A flag named value_cached is added to the Item_cache class. It's set to TRUE when cache holds the value of the last stored item. Function named cache_value() is added to the Item_cache class and derived classes. This function actually caches the value of the saved item. Item_cache_xxx::store functions now only store item to be cached and set value_cached flag to FALSE. Item_cache_xxx::val_xxx functions are changed to call cache_value function prior to returning cached value if value_cached is FALSE. The Arg_comparator::set_cmp_func function now calls cache_converted_constant to cache constants if they need a type conversion. The Item_cache::get_cache function is overloaded to allow setting of the cache type. The cache_converted_constant function is added to the Arg_comparator class. It checks whether a value can and should be cached and if so caches it.
2009-11-06 20:34:25 +01:00
Item** cache_converted_constant(THD *thd, Item **value, Item **cache,
Item_result type);
void set_datetime_cmp_func(Item_result_field *owner_arg, Item **a1, Item **b1);
2005-02-08 23:50:45 +01:00
static arg_cmp_func comparator_matrix [5][2];
inline bool is_owner_equal_func()
{
return (owner->type() == Item::FUNC_ITEM &&
((Item_func*)owner)->functype() == Item_func::EQUAL_FUNC);
}
void cleanup()
{
delete [] comparators;
comparators= 0;
}
Bug#49771: Incorrect MIN/MAX for date/time values. This bug is a design flaw of the fix for the bug#33546. It assumed that an item can be used only in one comparison context, but actually it isn't the case. Item_cache_datetime is used to store result for MIX/MAX aggregate functions. Because Arg_comparator always compares datetime values as INTs when possible the Item_cache_datetime most time caches only INT value. But since all datetime values has STRING result type MIN/MAX functions are asked for a STRING value when the result is being sent to a client. The Item_cache_datetime was designed to avoid conversions and get INT/STRING values from an underlying item, but at the moment the values is asked underlying item doesn't hold it anymore thus wrong result is returned. Beside that MIN/MAX aggregate functions was wrongly initializing cached result and this led to a wrong result. The Item::has_compatible_context helper function is added. It checks whether this and given items has the same comparison context or can be compared as DATETIME values by Arg_comparator. The equality propagation optimization is adjusted to take into account that items which being compared as DATETIME can have different comparison contexts. The Item_cache_datetime now converts cached INT value to a correct STRING DATETIME value by means of number_to_datetime & my_TIME_to_str functions. The Arg_comparator::set_cmp_context_for_datetime helper function is added. It sets comparison context of items being compared as DATETIMEs to INT if items will be compared as longlong. The Item_sum_hybrid::setup function now correctly initializes its result value. In order to avoid unnecessary conversions Item_sum_hybrid now states that it can provide correct longlong value if the item being aggregated can do it too.
2010-07-19 19:11:47 +02:00
/*
Set correct cmp_context if items would be compared as INTs.
*/
inline void set_cmp_context_for_datetime()
{
DBUG_ASSERT(func == &Arg_comparator::compare_datetime);
if ((*a)->result_as_longlong())
(*a)->cmp_context= INT_RESULT;
if ((*b)->result_as_longlong())
(*b)->cmp_context= INT_RESULT;
}
friend class Item_func;
};
2000-07-31 21:29:14 +02:00
class Item_bool_func :public Item_int_func
{
public:
Item_bool_func() :Item_int_func() {}
Item_bool_func(Item *a) :Item_int_func(a) {}
Item_bool_func(Item *a,Item *b) :Item_int_func(a,b) {}
Item_bool_func(THD *thd, Item_bool_func *item) :Item_int_func(thd, item) {}
2004-08-31 20:10:57 +02:00
bool is_bool_func() { return 1; }
2000-07-31 21:29:14 +02:00
void fix_length_and_dec() { decimals=0; max_length=1; }
uint decimal_precision() const { return 1; }
2000-07-31 21:29:14 +02:00
};
Bug#24532 (The return data type of IS TRUE is different from similar operations) Before this change, the boolean predicates: - X IS TRUE, - X IS NOT TRUE, - X IS FALSE, - X IS NOT FALSE were implemented by expanding the Item tree in the parser, by using a construct like: Item_func_if(Item_func_ifnull(X, <value>), <value>, <value>) Each <value> was a constant integer, either 0 or 1. A bug in the implementation of the function IF(a, b, c), in Item_func_if::fix_length_and_dec(), would cause the following : When the arguments b and c are both unsigned, the result type of the function was signed, instead of unsigned. When the result of the if function is signed, space for the sign could be counted twice (in the max() expression for a signed argument, and in the total), causing the member max_length to be too high. An effect of this is that the final type of IF(x, int(1), int(1)) would be int(2) instead of int(1). With this fix, the problems found in Item_func_if::fix_length_and_dec() have been fixed. While it's semantically correct to represent 'X IS TRUE' with Item_func_if(Item_func_ifnull(X, <value>), <value>, <value>), there are however more problems with this construct. a) Building the parse tree involves : - creating 5 Item instances (3 ints, 1 ifnull, 1 if), - creating each Item calls my_pthread_getspecific_ptr() once in the operator new(size), and a second time in the Item::Item() constructor, resulting in a total of 10 calls to get the current thread. Evaluating the expression involves evaluating up to 4 nodes at runtime. This representation could be greatly simplified and improved. b) Transforming the parse tree internally with if(ifnull(...)) is fine as long as this transformation is internal to the server implementation. With views however, the result of the parse tree is later exposed by the ::print() functions, and stored as part of the view definition. Doing this has long term consequences: 1) The original semantic 'X IS TRUE' is lost, and replaced by the if(ifnull(...)) expression. As a result, SHOW CREATE VIEW does not restore the original code. 2) Should a future version of MySQL implement the SQL BOOLEAN data type for example, views created today using 'X IS NULL' can be exported using mysqldump, and imported again. Such views would be converted correctly and automatically to use a BOOLEAN column in the future version. With 'X IS TRUE' and the current implementations, views using these "boolean" predicates would not be converted during the export/import, and would use integer columns instead. The difference traces back to how SHOW CREATE VIEW preserves 'X IS NULL' but does not preserve the 'X IS TRUE' semantic. With this fix, internal representation of 'X IS TRUE' booleans predicates has changed, so that: - dedicated Item classes are created for each predicate, - only 1 Item is created to represent 1 predicate - my_pthread_getspecific_ptr() is invoked 1 time instead of 10 - SHOW CREATE VIEW preserves the original semantic, and prints 'X IS TRUE'. Note that, because of the fix in Item_func_if, views created before this fix will: - correctly use a int(1) type instead of int(2) for boolean predicates, - incorrectly print the if(ifnull(...), ...) expression in SHOW CREATE VIEW, since the original semantic (X IS TRUE) has been lost. - except for the syntax used in SHOW CREATE VIEW, these views will operate properly, no action is needed. Views created after this fix will operate correctly, and will preserve the original code semantic in SHOW CREATE VIEW.
2007-02-12 21:59:29 +01:00
/**
Abstract Item class, to represent <code>X IS [NOT] (TRUE | FALSE)</code>
boolean predicates.
*/
class Item_func_truth : public Item_bool_func
{
public:
virtual bool val_bool();
virtual longlong val_int();
virtual void fix_length_and_dec();
virtual void print(String *str, enum_query_type query_type);
Bug#24532 (The return data type of IS TRUE is different from similar operations) Before this change, the boolean predicates: - X IS TRUE, - X IS NOT TRUE, - X IS FALSE, - X IS NOT FALSE were implemented by expanding the Item tree in the parser, by using a construct like: Item_func_if(Item_func_ifnull(X, <value>), <value>, <value>) Each <value> was a constant integer, either 0 or 1. A bug in the implementation of the function IF(a, b, c), in Item_func_if::fix_length_and_dec(), would cause the following : When the arguments b and c are both unsigned, the result type of the function was signed, instead of unsigned. When the result of the if function is signed, space for the sign could be counted twice (in the max() expression for a signed argument, and in the total), causing the member max_length to be too high. An effect of this is that the final type of IF(x, int(1), int(1)) would be int(2) instead of int(1). With this fix, the problems found in Item_func_if::fix_length_and_dec() have been fixed. While it's semantically correct to represent 'X IS TRUE' with Item_func_if(Item_func_ifnull(X, <value>), <value>, <value>), there are however more problems with this construct. a) Building the parse tree involves : - creating 5 Item instances (3 ints, 1 ifnull, 1 if), - creating each Item calls my_pthread_getspecific_ptr() once in the operator new(size), and a second time in the Item::Item() constructor, resulting in a total of 10 calls to get the current thread. Evaluating the expression involves evaluating up to 4 nodes at runtime. This representation could be greatly simplified and improved. b) Transforming the parse tree internally with if(ifnull(...)) is fine as long as this transformation is internal to the server implementation. With views however, the result of the parse tree is later exposed by the ::print() functions, and stored as part of the view definition. Doing this has long term consequences: 1) The original semantic 'X IS TRUE' is lost, and replaced by the if(ifnull(...)) expression. As a result, SHOW CREATE VIEW does not restore the original code. 2) Should a future version of MySQL implement the SQL BOOLEAN data type for example, views created today using 'X IS NULL' can be exported using mysqldump, and imported again. Such views would be converted correctly and automatically to use a BOOLEAN column in the future version. With 'X IS TRUE' and the current implementations, views using these "boolean" predicates would not be converted during the export/import, and would use integer columns instead. The difference traces back to how SHOW CREATE VIEW preserves 'X IS NULL' but does not preserve the 'X IS TRUE' semantic. With this fix, internal representation of 'X IS TRUE' booleans predicates has changed, so that: - dedicated Item classes are created for each predicate, - only 1 Item is created to represent 1 predicate - my_pthread_getspecific_ptr() is invoked 1 time instead of 10 - SHOW CREATE VIEW preserves the original semantic, and prints 'X IS TRUE'. Note that, because of the fix in Item_func_if, views created before this fix will: - correctly use a int(1) type instead of int(2) for boolean predicates, - incorrectly print the if(ifnull(...), ...) expression in SHOW CREATE VIEW, since the original semantic (X IS TRUE) has been lost. - except for the syntax used in SHOW CREATE VIEW, these views will operate properly, no action is needed. Views created after this fix will operate correctly, and will preserve the original code semantic in SHOW CREATE VIEW.
2007-02-12 21:59:29 +01:00
protected:
Item_func_truth(Item *a, bool a_value, bool a_affirmative)
: Item_bool_func(a), value(a_value), affirmative(a_affirmative)
{}
~Item_func_truth()
{}
private:
/**
True for <code>X IS [NOT] TRUE</code>,
false for <code>X IS [NOT] FALSE</code> predicates.
*/
const bool value;
/**
True for <code>X IS Y</code>, false for <code>X IS NOT Y</code> predicates.
*/
const bool affirmative;
};
/**
This Item represents a <code>X IS TRUE</code> boolean predicate.
*/
class Item_func_istrue : public Item_func_truth
{
public:
Item_func_istrue(Item *a) : Item_func_truth(a, true, true) {}
~Item_func_istrue() {}
virtual const char* func_name() const { return "istrue"; }
};
/**
This Item represents a <code>X IS NOT TRUE</code> boolean predicate.
*/
class Item_func_isnottrue : public Item_func_truth
{
public:
Item_func_isnottrue(Item *a) : Item_func_truth(a, true, false) {}
~Item_func_isnottrue() {}
virtual const char* func_name() const { return "isnottrue"; }
};
/**
This Item represents a <code>X IS FALSE</code> boolean predicate.
*/
class Item_func_isfalse : public Item_func_truth
{
public:
Item_func_isfalse(Item *a) : Item_func_truth(a, false, true) {}
~Item_func_isfalse() {}
virtual const char* func_name() const { return "isfalse"; }
};
/**
This Item represents a <code>X IS NOT FALSE</code> boolean predicate.
*/
class Item_func_isnotfalse : public Item_func_truth
{
public:
Item_func_isnotfalse(Item *a) : Item_func_truth(a, false, false) {}
~Item_func_isnotfalse() {}
virtual const char* func_name() const { return "isnotfalse"; }
};
class Item_cache;
#define UNKNOWN ((my_bool)-1)
/*
Item_in_optimizer(left_expr, Item_in_subselect(...))
Item_in_optimizer is used to wrap an instance of Item_in_subselect. This
class does the following:
- Evaluate the left expression and store it in Item_cache_* object (to
avoid re-evaluating it many times during subquery execution)
- Shortcut the evaluation of "NULL IN (...)" to NULL in the cases where we
don't care if the result is NULL or FALSE.
NOTE
It is not quite clear why the above listed functionality should be
placed into a separate class called 'Item_in_optimizer'.
*/
class Item_in_optimizer: public Item_bool_func
{
protected:
Item_cache *cache;
bool save_cache;
/*
Stores the value of "NULL IN (SELECT ...)" for uncorrelated subqueries:
UNKNOWN - "NULL in (SELECT ...)" has not yet been evaluated
FALSE - result is FALSE
TRUE - result is NULL
*/
my_bool result_for_null_param;
public:
2002-12-19 20:15:09 +01:00
Item_in_optimizer(Item *a, Item_in_subselect *b):
Item_bool_func(a, reinterpret_cast<Item *>(b)), cache(0),
save_cache(0), result_for_null_param(UNKNOWN)
{ with_subselect= true; }
bool fix_fields(THD *, Item **);
bool fix_left(THD *thd, Item **ref);
bool is_null();
longlong val_int();
void cleanup();
const char *func_name() const { return "<in_optimizer>"; }
2002-12-19 20:15:09 +01:00
Item_cache **get_cache() { return &cache; }
void keep_top_level_cache();
};
2003-11-03 11:28:36 +01:00
class Comp_creator
{
public:
Comp_creator() {} /* Remove gcc warning */
virtual ~Comp_creator() {} /* Remove gcc warning */
2003-11-03 11:28:36 +01:00
virtual Item_bool_func2* create(Item *a, Item *b) const = 0;
virtual const char* symbol(bool invert) const = 0;
virtual bool eqne_op() const = 0;
virtual bool l_op() const = 0;
};
class Eq_creator :public Comp_creator
{
public:
Eq_creator() {} /* Remove gcc warning */
virtual ~Eq_creator() {} /* Remove gcc warning */
2003-11-03 11:28:36 +01:00
virtual Item_bool_func2* create(Item *a, Item *b) const;
virtual const char* symbol(bool invert) const { return invert? "<>" : "="; }
virtual bool eqne_op() const { return 1; }
virtual bool l_op() const { return 0; }
};
class Ne_creator :public Comp_creator
{
public:
Ne_creator() {} /* Remove gcc warning */
virtual ~Ne_creator() {} /* Remove gcc warning */
2003-11-03 11:28:36 +01:00
virtual Item_bool_func2* create(Item *a, Item *b) const;
virtual const char* symbol(bool invert) const { return invert? "=" : "<>"; }
virtual bool eqne_op() const { return 1; }
virtual bool l_op() const { return 0; }
};
class Gt_creator :public Comp_creator
{
public:
Gt_creator() {} /* Remove gcc warning */
virtual ~Gt_creator() {} /* Remove gcc warning */
2003-11-03 11:28:36 +01:00
virtual Item_bool_func2* create(Item *a, Item *b) const;
virtual const char* symbol(bool invert) const { return invert? "<=" : ">"; }
virtual bool eqne_op() const { return 0; }
virtual bool l_op() const { return 0; }
};
class Lt_creator :public Comp_creator
{
public:
Lt_creator() {} /* Remove gcc warning */
virtual ~Lt_creator() {} /* Remove gcc warning */
2003-11-03 11:28:36 +01:00
virtual Item_bool_func2* create(Item *a, Item *b) const;
virtual const char* symbol(bool invert) const { return invert? ">=" : "<"; }
virtual bool eqne_op() const { return 0; }
virtual bool l_op() const { return 1; }
};
class Ge_creator :public Comp_creator
{
public:
Ge_creator() {} /* Remove gcc warning */
virtual ~Ge_creator() {} /* Remove gcc warning */
2003-11-03 11:28:36 +01:00
virtual Item_bool_func2* create(Item *a, Item *b) const;
virtual const char* symbol(bool invert) const { return invert? "<" : ">="; }
virtual bool eqne_op() const { return 0; }
virtual bool l_op() const { return 0; }
};
class Le_creator :public Comp_creator
{
public:
Le_creator() {} /* Remove gcc warning */
virtual ~Le_creator() {} /* Remove gcc warning */
2003-11-03 11:28:36 +01:00
virtual Item_bool_func2* create(Item *a, Item *b) const;
virtual const char* symbol(bool invert) const { return invert? ">" : "<="; }
virtual bool eqne_op() const { return 0; }
virtual bool l_op() const { return 1; }
};
2000-07-31 21:29:14 +02:00
class Item_bool_func2 :public Item_int_func
{ /* Bool with 2 string args */
protected:
Arg_comparator cmp;
bool abort_on_null;
2000-07-31 21:29:14 +02:00
public:
Item_bool_func2(Item *a,Item *b)
:Item_int_func(a,b), cmp(tmp_arg, tmp_arg+1), abort_on_null(FALSE) {}
2000-07-31 21:29:14 +02:00
void fix_length_and_dec();
void set_cmp_func()
2002-11-15 19:32:09 +01:00
{
cmp.set_cmp_func(this, tmp_arg, tmp_arg+1, TRUE);
2002-11-15 19:32:09 +01:00
}
2000-07-31 21:29:14 +02:00
optimize_type select_optimize() const { return OPTIMIZE_OP; }
virtual enum Functype rev_functype() const { return UNKNOWN_FUNC; }
bool have_rev_func() const { return rev_functype() != UNKNOWN_FUNC; }
virtual inline void print(String *str, enum_query_type query_type)
{
Item_func::print_op(str, query_type);
}
bool is_null() { return test(args[0]->is_null() || args[1]->is_null()); }
2004-08-31 20:10:57 +02:00
bool is_bool_func() { return 1; }
CHARSET_INFO *compare_collation() { return cmp.cmp_collation.collation; }
uint decimal_precision() const { return 1; }
void top_level_item() { abort_on_null= TRUE; }
void cleanup()
{
Item_int_func::cleanup();
cmp.cleanup();
}
friend class Arg_comparator;
2000-07-31 21:29:14 +02:00
};
2002-11-15 19:32:09 +01:00
class Item_bool_rowready_func2 :public Item_bool_func2
{
public:
A fix and test case for Bug#5987 "subselect in bool function crashes server (prepared statements)": the bug was that all boolean items always recovered its original arguments at statement cleanup stage. This collided with Item_subselect::select_transformer, which tries to permanently change the item tree to use a transformed subselect instead of original one. So we had this call sequence for prepare: mysql_stmt_prepare -> JOIN::prepare -> Item_subselect::fix_fields -> the item tree gets transformed -> Item_bool_rowready_func2::cleanup, item tree is recovered to original state, while it shouldn't have been; mysql_stmt_execute -> attempts to execute a broken tree -> crash. Now instead of bluntly recovering all arguments of bool functions in Item_bool_rowready_func2::cleanup, we recover only those which were changed, and do it in one place. There still would exist a possibility for a collision with subselect tranformation, if permanent and temporary changes were performed at the same stage. But fortunately subselect transformation is always done first, so it doesn't conflict with the optimization done by propogate_cond_constants. Now we have: mysql_stmt_prepare -> JOIN::prepare -> subselect transformation permanently changes the tree -> cleanup doesn't recover anything, because nothing was registered for recovery. mysql_stmt_execute -> JOIN::prepare (the tree is already transformed, so it doesn't change), JOIN::optimize -> propogate_cond_constants -> temporary changes the item tree with constants -> JOIN::execute -> cleanup -> the changes done by propogate_cond_constants are recovered, as they were registered for recovery.
2004-10-10 00:39:22 +02:00
Item_bool_rowready_func2(Item *a, Item *b) :Item_bool_func2(a, b)
2002-11-15 19:32:09 +01:00
{
allowed_arg_cols= 0; // Fetch this value from first argument
2002-11-15 19:32:09 +01:00
}
Item *neg_transformer(THD *thd);
virtual Item *negated_item();
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
bool subst_argument_checker(uchar **arg) { return TRUE; }
2002-11-15 19:32:09 +01:00
};
2000-07-31 21:29:14 +02:00
class Item_func_not :public Item_bool_func
{
public:
Item_func_not(Item *a) :Item_bool_func(a) {}
longlong val_int();
enum Functype functype() const { return NOT_FUNC; }
2000-07-31 21:29:14 +02:00
const char *func_name() const { return "not"; }
Item *neg_transformer(THD *thd);
virtual void print(String *str, enum_query_type query_type);
2000-07-31 21:29:14 +02:00
};
class Item_maxmin_subselect;
/*
trigcond<param>(arg) ::= param? arg : TRUE
The class Item_func_trig_cond is used for guarded predicates
which are employed only for internal purposes.
A guarded predicate is an object consisting of an a regular or
a guarded predicate P and a pointer to a boolean guard variable g.
A guarded predicate P/g is evaluated to true if the value of the
guard g is false, otherwise it is evaluated to the same value that
the predicate P: val(P/g)= g ? val(P):true.
Guarded predicates allow us to include predicates into a conjunction
conditionally. Currently they are utilized for pushed down predicates
in queries with outer join operations.
In the future, probably, it makes sense to extend this class to
the objects consisting of three elements: a predicate P, a pointer
to a variable g and a firing value s with following evaluation
rule: val(P/g,s)= g==s? val(P) : true. It will allow us to build only
one item for the objects of the form P/g1/g2...
Objects of this class are built only for query execution after
the execution plan has been already selected. That's why this
class needs only val_int out of generic methods.
Current uses of Item_func_trig_cond objects:
- To wrap selection conditions when executing outer joins
- To wrap condition that is pushed down into subquery
*/
class Item_func_trig_cond: public Item_bool_func
{
bool *trig_var;
public:
Item_func_trig_cond(Item *a, bool *f) : Item_bool_func(a) { trig_var= f; }
longlong val_int() { return *trig_var ? args[0]->val_int() : 1; }
enum Functype functype() const { return TRIG_COND_FUNC; };
const char *func_name() const { return "trigcond"; };
bool const_item() const { return FALSE; }
bool *get_trig_var() { return trig_var; }
};
class Item_func_not_all :public Item_func_not
{
2005-02-08 23:50:45 +01:00
/* allow to check presence of values in max/min optimization */
Item_sum_hybrid *test_sum_item;
Item_maxmin_subselect *test_sub_item;
bool abort_on_null;
public:
bool show;
Item_func_not_all(Item *a)
:Item_func_not(a), test_sum_item(0), test_sub_item(0), abort_on_null(0),
show(0)
{}
virtual void top_level_item() { abort_on_null= 1; }
bool top_level() { return abort_on_null; }
longlong val_int();
enum Functype functype() const { return NOT_ALL_FUNC; }
const char *func_name() const { return "<not>"; }
virtual void print(String *str, enum_query_type query_type);
void set_sum_test(Item_sum_hybrid *item) { test_sum_item= item; };
void set_sub_test(Item_maxmin_subselect *item) { test_sub_item= item; };
bool empty_underlying_subquery();
Item *neg_transformer(THD *thd);
};
class Item_func_nop_all :public Item_func_not_all
{
public:
Item_func_nop_all(Item *a) :Item_func_not_all(a) {}
longlong val_int();
const char *func_name() const { return "<nop>"; }
Item *neg_transformer(THD *thd);
};
2002-11-15 19:32:09 +01:00
class Item_func_eq :public Item_bool_rowready_func2
2000-07-31 21:29:14 +02:00
{
public:
Item_func_eq(Item *a,Item *b) :Item_bool_rowready_func2(a,b) {}
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return EQ_FUNC; }
enum Functype rev_functype() const { return EQ_FUNC; }
cond_result eq_cmp_result() const { return COND_TRUE; }
const char *func_name() const { return "="; }
Item *negated_item();
2000-07-31 21:29:14 +02:00
};
2002-11-28 08:11:35 +01:00
class Item_func_equal :public Item_bool_rowready_func2
2000-07-31 21:29:14 +02:00
{
public:
2002-11-28 19:58:23 +01:00
Item_func_equal(Item *a,Item *b) :Item_bool_rowready_func2(a,b) {};
2000-07-31 21:29:14 +02:00
longlong val_int();
void fix_length_and_dec();
table_map not_null_tables() const { return 0; }
2000-07-31 21:29:14 +02:00
enum Functype functype() const { return EQUAL_FUNC; }
enum Functype rev_functype() const { return EQUAL_FUNC; }
cond_result eq_cmp_result() const { return COND_TRUE; }
const char *func_name() const { return "<=>"; }
Item *neg_transformer(THD *thd) { return 0; }
2000-07-31 21:29:14 +02:00
};
2002-11-15 19:32:09 +01:00
class Item_func_ge :public Item_bool_rowready_func2
2000-07-31 21:29:14 +02:00
{
public:
Item_func_ge(Item *a,Item *b) :Item_bool_rowready_func2(a,b) {};
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return GE_FUNC; }
enum Functype rev_functype() const { return LE_FUNC; }
cond_result eq_cmp_result() const { return COND_TRUE; }
const char *func_name() const { return ">="; }
Item *negated_item();
2000-07-31 21:29:14 +02:00
};
2002-11-15 19:32:09 +01:00
class Item_func_gt :public Item_bool_rowready_func2
2000-07-31 21:29:14 +02:00
{
public:
Item_func_gt(Item *a,Item *b) :Item_bool_rowready_func2(a,b) {};
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return GT_FUNC; }
enum Functype rev_functype() const { return LT_FUNC; }
cond_result eq_cmp_result() const { return COND_FALSE; }
const char *func_name() const { return ">"; }
Item *negated_item();
2000-07-31 21:29:14 +02:00
};
2002-11-15 19:32:09 +01:00
class Item_func_le :public Item_bool_rowready_func2
2000-07-31 21:29:14 +02:00
{
public:
Item_func_le(Item *a,Item *b) :Item_bool_rowready_func2(a,b) {};
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return LE_FUNC; }
enum Functype rev_functype() const { return GE_FUNC; }
cond_result eq_cmp_result() const { return COND_TRUE; }
const char *func_name() const { return "<="; }
Item *negated_item();
2000-07-31 21:29:14 +02:00
};
2002-11-15 19:32:09 +01:00
class Item_func_lt :public Item_bool_rowready_func2
2000-07-31 21:29:14 +02:00
{
public:
Item_func_lt(Item *a,Item *b) :Item_bool_rowready_func2(a,b) {}
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return LT_FUNC; }
enum Functype rev_functype() const { return GT_FUNC; }
cond_result eq_cmp_result() const { return COND_FALSE; }
const char *func_name() const { return "<"; }
Item *negated_item();
2000-07-31 21:29:14 +02:00
};
2002-11-15 19:32:09 +01:00
class Item_func_ne :public Item_bool_rowready_func2
2000-07-31 21:29:14 +02:00
{
public:
Item_func_ne(Item *a,Item *b) :Item_bool_rowready_func2(a,b) {}
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return NE_FUNC; }
cond_result eq_cmp_result() const { return COND_FALSE; }
optimize_type select_optimize() const { return OPTIMIZE_KEY; }
2000-07-31 21:29:14 +02:00
const char *func_name() const { return "<>"; }
Item *negated_item();
2000-07-31 21:29:14 +02:00
};
/*
The class Item_func_opt_neg is defined to factor out the functionality
common for the classes Item_func_between and Item_func_in. The objects
of these classes can express predicates or there negations.
The alternative approach would be to create pairs Item_func_between,
Item_func_notbetween and Item_func_in, Item_func_notin.
*/
class Item_func_opt_neg :public Item_int_func
{
public:
bool negated; /* <=> the item represents NOT <func> */
bool pred_level; /* <=> [NOT] <func> is used on a predicate level */
public:
Item_func_opt_neg(Item *a, Item *b, Item *c)
:Item_int_func(a, b, c), negated(0), pred_level(0) {}
Item_func_opt_neg(List<Item> &list)
:Item_int_func(list), negated(0), pred_level(0) {}
public:
inline void negate() { negated= !negated; }
inline void top_level_item() { pred_level= 1; }
Item *neg_transformer(THD *thd)
{
negated= !negated;
return this;
}
bool eq(const Item *item, bool binary_cmp) const;
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
bool subst_argument_checker(uchar **arg) { return TRUE; }
};
class Item_func_between :public Item_func_opt_neg
2000-07-31 21:29:14 +02:00
{
DTCollation cmp_collation;
2000-07-31 21:29:14 +02:00
public:
Item_result cmp_type;
String value0,value1,value2;
/* TRUE <=> arguments will be compared as dates. */
bool compare_as_dates;
/* Comparators used for DATE/DATETIME comparison. */
Arg_comparator ge_cmp, le_cmp;
Item_func_between(Item *a, Item *b, Item *c)
:Item_func_opt_neg(a, b, c), compare_as_dates(FALSE) {}
2000-07-31 21:29:14 +02:00
longlong val_int();
optimize_type select_optimize() const { return OPTIMIZE_KEY; }
enum Functype functype() const { return BETWEEN; }
const char *func_name() const { return "between"; }
bool fix_fields(THD *, Item **);
2000-07-31 21:29:14 +02:00
void fix_length_and_dec();
virtual void print(String *str, enum_query_type query_type);
bool is_bool_func() { return 1; }
CHARSET_INFO *compare_collation() { return cmp_collation.collation; }
uint decimal_precision() const { return 1; }
2000-07-31 21:29:14 +02:00
};
class Item_func_strcmp :public Item_bool_func2
{
public:
Item_func_strcmp(Item *a,Item *b) :Item_bool_func2(a,b) {}
longlong val_int();
optimize_type select_optimize() const { return OPTIMIZE_NONE; }
const char *func_name() const { return "strcmp"; }
virtual inline void print(String *str, enum_query_type query_type)
{
Item_func::print(str, query_type);
}
WL#2649 Number-to-string conversions added: include/ctype_numconv.inc mysql-test/include/ctype_numconv.inc mysql-test/r/ctype_binary.result mysql-test/t/ctype_binary.test Adding tests modified: mysql-test/r/bigint.result mysql-test/r/case.result mysql-test/r/create.result mysql-test/r/ctype_cp1251.result mysql-test/r/ctype_latin1.result mysql-test/r/ctype_ucs.result mysql-test/r/func_gconcat.result mysql-test/r/func_str.result mysql-test/r/metadata.result mysql-test/r/ps_1general.result mysql-test/r/ps_2myisam.result mysql-test/r/ps_3innodb.result mysql-test/r/ps_4heap.result mysql-test/r/ps_5merge.result mysql-test/r/show_check.result mysql-test/r/type_datetime.result mysql-test/r/type_ranges.result mysql-test/r/union.result mysql-test/suite/ndb/r/ps_7ndb.result mysql-test/t/ctype_cp1251.test mysql-test/t/ctype_latin1.test mysql-test/t/ctype_ucs.test mysql-test/t/func_str.test Fixing tests @ sql/field.cc - Return str result using my_charset_numeric. - Using real multi-byte aware str_to_XXX functions to handle tricky charset values propely (e.g. UCS2) @ sql/field.h - Changing derivation of non-string field types to DERIVATION_NUMERIC. - Changing binary() for numeric/datetime fields to always return TRUE even if charset is not my_charset_bin. We need this to keep ha_base_keytype() return HA_KEYTYPE_BINARY. - Adding BINARY_FLAG into some fields, because it's not being set automatically anymore with "my_charset_bin to my_charset_numeric" change. - Changing derivation for numeric/datetime datatypes to a weaker value, to make "SELECT concat('string', field)" use character set of the string literal for the result of the function. @ sql/item.cc - Implementing generic val_str_ascii(). - Using max_char_length() instead of direct read of max_length to make "tricky" charsets like UCS2 work. NOTE: in the future we'll possibly remove all direct reads of max_length - Fixing Item_num::safe_charset_converter(). Previously it alligned binary string to character string (for example by adding leading 0x00 when doing binary->UCS2 conversion). Now it just converts from my_charset_numbner to "tocs". - Using val_str_ascii() in Item::get_time() to make UCS2 arguments work. - Other misc changes @ sql/item.h - Changing MY_COLL_CMP_CONV and MY_COLL_ALLOW_CONV to bit operations instead of hard-coded bit masks. - Addding new method DTCollation.set_numeric(). - Adding new methods to Item. - Adding helper functions to make code look nicer: agg_item_charsets_for_string_result() agg_item_charsets_for_comparison() - Changing charset for Item_num-derived items from my_charset_bin to my_charset_numeric (which is an alias for latin1). @ sql/item_cmpfunc.cc - Using new helper functions - Other misc changes @ sql/item_cmpfunc.h - Fixing strcmp() to return max_length=2. Previously it returned 1, which was wrong, because it did not fit '-1'. @ sql/item_func.cc - Using new helper functions - Other minor changes @ sql/item_func.h - Removing unused functions - Adding helper functions agg_arg_charsets_for_string_result() agg_arg_charsets_for_comparison() - Adding set_numeric() into constructors of numeric items. - Using fix_length_and_charset() and fix_char_length() instead of direct write to max_length. @ sql/item_geofunc.cc - Changing class for Item_func_geometry_type and Item_func_as_wkt from Item_str_func to Item_str_ascii_func, to make them return UCS2 result properly (when character_set_connection=ucs2). @ sql/item_geofunc.h - Changing class for Item_func_geometry_type and Item_func_as_wkt from Item_str_func to Item_str_ascii_func, to make them return UCS2 result properly (when @@character_set_connection=ucs2). @ sql/item_strfunc.cc - Implementing Item_str_func::val_str(). - Renaming val_str to val_str_ascii for some items, to make them work with UCS2 properly. - Using new helper functions - All single-argument functions that expect string result now call this method: agg_arg_charsets_for_string_result(collation, args, 1); This enables character set conversion to @@character_set_connection in case of pure numeric input. @ sql/item_strfunc.h - Introducing Item_str_ascii_func - for functions which return pure ASCII data, for performance purposes, as well as for the cases when the old implementation of val_str() was heavily 8-bit oriented and implementing a UCS2-aware version is tricky. @ sql/item_sum.cc - Using new helper functions. @ sql/item_timefunc.cc - Using my_charset_numeric instead of my_charset_bin. - Using fix_char_length(), fix_length_and_charset() and fix_length_and_charset_datetime() instead of direct write to max_length. - Using tricky-charset aware function str_to_time_with_warn() @ sql/item_timefunc.h - Using new helper functions for charset and length initialization. - Changing base class for Item_func_get_format() to make it return UCS2 properly (when character_set_connection=ucs2). @ sql/item_xmlfunc.cc - Using new helper function @ sql/my_decimal.cc - Adding a new DECIMAL to CHAR converter with real multibyte support (e.g. UCS2) @ sql/mysql_priv.h - Introducing a new derivation level for numeric/datetime data types. - Adding macros for my_charset_numeric and MY_REPERTOIRE_NUMERIC. - Adding prototypes for str_set_decimal() - Adding prototypes for character-set aware str_to_xxx() functions. @ sql/protocol.cc - Changing charsetnr to "binary" client-side metadata for numeric/datetime data types. @ sql/time.cc - Adding to_ascii() helper function, to convert a string in any character set to ascii representation. In the future can be extended to understand digits written in various non-Latin word scripts. - Adding real multy-byte character set aware versions for str_to_XXXX, to make these these type of queries work correct: INSERT INTO t1 SET datetime_column=ucs2_expression; @ strings/ctype-ucs2.c - endptr was not calculated correctly. INSERTing of UCS2 values into numeric columns returned warnings about truncated wrong data.
2010-02-11 05:17:25 +01:00
void fix_length_and_dec()
{
Item_bool_func2::fix_length_and_dec();
fix_char_length(2); // returns "1" or "0" or "-1"
}
2000-07-31 21:29:14 +02:00
};
2005-02-08 23:50:45 +01:00
struct interval_range
{
Item_result type;
double dbl;
my_decimal dec;
};
2000-07-31 21:29:14 +02:00
class Item_func_interval :public Item_int_func
{
Item_row *row;
2005-02-08 23:50:45 +01:00
my_bool use_decimal_comparison;
interval_range *intervals;
2000-07-31 21:29:14 +02:00
public:
Item_func_interval(Item_row *a)
:Item_int_func(a),row(a),intervals(0)
{
allowed_arg_cols= 0; // Fetch this value from first argument
}
2000-07-31 21:29:14 +02:00
longlong val_int();
void fix_length_and_dec();
const char *func_name() const { return "interval"; }
uint decimal_precision() const { return 2; }
2000-07-31 21:29:14 +02:00
};
class Item_func_coalesce :public Item_func_numhybrid
2000-07-31 21:29:14 +02:00
{
2005-02-08 23:50:45 +01:00
protected:
enum_field_types cached_field_type;
Item_func_coalesce(Item *a, Item *b) :Item_func_numhybrid(a, b) {}
2005-02-08 23:50:45 +01:00
public:
Item_func_coalesce(List<Item> &list) :Item_func_numhybrid(list) {}
double real_op();
longlong int_op();
String *str_op(String *);
my_decimal *decimal_op(my_decimal *);
2005-02-08 23:50:45 +01:00
void fix_length_and_dec();
void find_num_type() {}
enum Item_result result_type () const { return hybrid_type; }
2005-02-08 23:50:45 +01:00
const char *func_name() const { return "coalesce"; }
table_map not_null_tables() const { return 0; }
enum_field_types field_type() const { return cached_field_type; }
2005-02-08 23:50:45 +01:00
};
class Item_func_ifnull :public Item_func_coalesce
{
protected:
bool field_type_defined;
2000-07-31 21:29:14 +02:00
public:
2005-02-08 23:50:45 +01:00
Item_func_ifnull(Item *a, Item *b) :Item_func_coalesce(a,b) {}
double real_op();
longlong int_op();
String *str_op(String *str);
my_decimal *decimal_op(my_decimal *);
enum_field_types field_type() const;
2000-07-31 21:29:14 +02:00
void fix_length_and_dec();
const char *func_name() const { return "ifnull"; }
Field *tmp_table_field(TABLE *table);
uint decimal_precision() const;
2000-07-31 21:29:14 +02:00
};
class Item_func_if :public Item_func
{
enum Item_result cached_result_type;
enum_field_types cached_field_type;
2000-07-31 21:29:14 +02:00
public:
Item_func_if(Item *a,Item *b,Item *c)
:Item_func(a,b,c), cached_result_type(INT_RESULT)
{}
2004-11-11 19:39:35 +01:00
double val_real();
2000-07-31 21:29:14 +02:00
longlong val_int();
String *val_str(String *str);
2005-02-08 23:50:45 +01:00
my_decimal *val_decimal(my_decimal *);
2000-07-31 21:29:14 +02:00
enum Item_result result_type () const { return cached_result_type; }
enum_field_types field_type() const { return cached_field_type; }
bool fix_fields(THD *, Item **);
2000-07-31 21:29:14 +02:00
void fix_length_and_dec();
uint decimal_precision() const;
2000-07-31 21:29:14 +02:00
const char *func_name() const { return "if"; }
private:
void cache_type_info(Item *source);
2000-07-31 21:29:14 +02:00
};
class Item_func_nullif :public Item_bool_func2
{
enum Item_result cached_result_type;
public:
Item_func_nullif(Item *a,Item *b)
:Item_bool_func2(a,b), cached_result_type(INT_RESULT)
{}
2004-11-11 19:39:35 +01:00
double val_real();
2000-07-31 21:29:14 +02:00
longlong val_int();
String *val_str(String *str);
2005-02-08 23:50:45 +01:00
my_decimal *val_decimal(my_decimal *);
2000-07-31 21:29:14 +02:00
enum Item_result result_type () const { return cached_result_type; }
void fix_length_and_dec();
uint decimal_precision() const { return args[0]->decimal_precision(); }
2000-07-31 21:29:14 +02:00
const char *func_name() const { return "nullif"; }
virtual inline void print(String *str, enum_query_type query_type)
{
Item_func::print(str, query_type);
}
table_map not_null_tables() const { return 0; }
bool is_null();
2000-07-31 21:29:14 +02:00
};
/* Functions to handle the optimized IN */
/* A vector of values of some type */
2000-07-31 21:29:14 +02:00
class in_vector :public Sql_alloc
{
public:
2000-07-31 21:29:14 +02:00
char *base;
uint size;
qsort2_cmp compare;
CHARSET_INFO *collation;
2000-07-31 21:29:14 +02:00
uint count;
uint used_count;
in_vector() {}
in_vector(uint elements,uint element_length,qsort2_cmp cmp_func,
CHARSET_INFO *cmp_coll)
2000-07-31 21:29:14 +02:00
:base((char*) sql_calloc(elements*element_length)),
size(element_length), compare(cmp_func), collation(cmp_coll),
count(elements), used_count(elements) {}
2000-07-31 21:29:14 +02:00
virtual ~in_vector() {}
virtual void set(uint pos,Item *item)=0;
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
virtual uchar *get_value(Item *item)=0;
2000-07-31 21:29:14 +02:00
void sort()
{
my_qsort2(base,used_count,size,compare,collation);
}
2000-07-31 21:29:14 +02:00
int find(Item *item);
/*
Create an instance of Item_{type} (e.g. Item_decimal) constant object
which type allows it to hold an element of this vector without any
conversions.
The purpose of this function is to be able to get elements of this
vector in form of Item_xxx constants without creating Item_xxx object
for every array element you get (i.e. this implements "FlyWeight" pattern)
*/
virtual Item* create_item() { return NULL; }
/*
Store the value at position #pos into provided item object
SYNOPSIS
value_to_item()
pos Index of value to store
item Constant item to store value into. The item must be of the same
type that create_item() returns.
*/
virtual void value_to_item(uint pos, Item *item) { }
/* Compare values number pos1 and pos2 for equality */
bool compare_elems(uint pos1, uint pos2)
{
return test(compare(collation, base + pos1*size, base + pos2*size));
}
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
virtual Item_result result_type()= 0;
2000-07-31 21:29:14 +02:00
};
class in_string :public in_vector
{
2005-02-08 23:50:45 +01:00
char buff[STRING_BUFFER_USUAL_SIZE];
2000-07-31 21:29:14 +02:00
String tmp;
public:
in_string(uint elements,qsort2_cmp cmp_func, CHARSET_INFO *cs);
2000-07-31 21:29:14 +02:00
~in_string();
void set(uint pos,Item *item);
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
uchar *get_value(Item *item);
Item* create_item()
{
return new Item_string(collation);
}
void value_to_item(uint pos, Item *item)
{
String *str=((String*) base)+pos;
Item_string *to= (Item_string*)item;
to->str_value= *str;
}
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
Item_result result_type() { return STRING_RESULT; }
2000-07-31 21:29:14 +02:00
};
class in_longlong :public in_vector
{
protected:
/*
Here we declare a temporary variable (tmp) of the same type as the
elements of this vector. tmp is used in finding if a given value is in
the list.
*/
struct packed_longlong
{
longlong val;
longlong unsigned_flag; // Use longlong, not bool, to preserve alignment
} tmp;
2000-07-31 21:29:14 +02:00
public:
in_longlong(uint elements);
void set(uint pos,Item *item);
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
uchar *get_value(Item *item);
Item* create_item()
{
/*
We're created a signed INT, this may not be correct in
general case (see BUG#19342).
*/
2006-04-25 23:21:33 +02:00
return new Item_int((longlong)0);
}
void value_to_item(uint pos, Item *item)
{
((Item_int*) item)->value= ((packed_longlong*) base)[pos].val;
2007-03-02 18:32:46 +01:00
((Item_int*) item)->unsigned_flag= (my_bool)
((packed_longlong*) base)[pos].unsigned_flag;
}
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
Item_result result_type() { return INT_RESULT; }
friend int cmp_longlong(void *cmp_arg, packed_longlong *a,packed_longlong *b);
2000-07-31 21:29:14 +02:00
};
/*
Class to represent a vector of constant DATE/DATETIME values.
Values are obtained with help of the get_datetime_value() function.
If the left item is a constant one then its value is cached in the
lval_cache variable.
*/
class in_datetime :public in_longlong
{
public:
THD *thd;
/* An item used to issue warnings. */
Item *warn_item;
/* Cache for the left item. */
Item *lval_cache;
in_datetime(Item *warn_item_arg, uint elements)
:in_longlong(elements), thd(current_thd), warn_item(warn_item_arg),
lval_cache(0) {};
void set(uint pos,Item *item);
2007-05-24 18:47:58 +02:00
uchar *get_value(Item *item);
friend int cmp_longlong(void *cmp_arg, packed_longlong *a,packed_longlong *b);
};
2000-07-31 21:29:14 +02:00
class in_double :public in_vector
{
double tmp;
public:
in_double(uint elements);
void set(uint pos,Item *item);
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
uchar *get_value(Item *item);
Item *create_item()
{
return new Item_float(0.0, 0);
}
void value_to_item(uint pos, Item *item)
{
((Item_float*)item)->value= ((double*) base)[pos];
}
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
Item_result result_type() { return REAL_RESULT; }
2000-07-31 21:29:14 +02:00
};
2005-02-08 23:50:45 +01:00
class in_decimal :public in_vector
{
my_decimal val;
public:
in_decimal(uint elements);
void set(uint pos, Item *item);
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
uchar *get_value(Item *item);
Item *create_item()
{
return new Item_decimal(0, FALSE);
}
void value_to_item(uint pos, Item *item)
{
my_decimal *dec= ((my_decimal *)base) + pos;
Item_decimal *item_dec= (Item_decimal*)item;
item_dec->set_decimal_value(dec);
}
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
Item_result result_type() { return DECIMAL_RESULT; }
2005-02-08 23:50:45 +01:00
};
2000-07-31 21:29:14 +02:00
/*
** Classes for easy comparing of non const items
*/
class cmp_item :public Sql_alloc
{
public:
CHARSET_INFO *cmp_charset;
cmp_item() { cmp_charset= &my_charset_bin; }
2000-07-31 21:29:14 +02:00
virtual ~cmp_item() {}
virtual void store_value(Item *item)= 0;
virtual int cmp(Item *item)= 0;
// for optimized IN with row
virtual int compare(cmp_item *item)= 0;
2005-02-08 23:50:45 +01:00
static cmp_item* get_comparator(Item_result type, CHARSET_INFO *cs);
virtual cmp_item *make_same()= 0;
virtual void store_value_by_template(cmp_item *tmpl, Item *item)
{
store_value(item);
}
2000-07-31 21:29:14 +02:00
};
class cmp_item_string :public cmp_item
{
protected:
String *value_res;
public:
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
cmp_item_string () {}
cmp_item_string (CHARSET_INFO *cs) { cmp_charset= cs; }
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
void set_charset(CHARSET_INFO *cs) { cmp_charset= cs; }
friend class cmp_item_sort_string;
friend class cmp_item_sort_string_in_static;
};
2000-07-31 21:29:14 +02:00
class cmp_item_sort_string :public cmp_item_string
{
protected:
2005-02-08 23:50:45 +01:00
char value_buff[STRING_BUFFER_USUAL_SIZE];
String value;
2000-07-31 21:29:14 +02:00
public:
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
cmp_item_sort_string():
cmp_item_string() {}
cmp_item_sort_string(CHARSET_INFO *cs):
cmp_item_string(cs),
value(value_buff, sizeof(value_buff), cs) {}
2000-07-31 21:29:14 +02:00
void store_value(Item *item)
{
value_res= item->val_str(&value);
}
2000-07-31 21:29:14 +02:00
int cmp(Item *arg)
{
2005-02-08 23:50:45 +01:00
char buff[STRING_BUFFER_USUAL_SIZE];
String tmp(buff, sizeof(buff), cmp_charset), *res;
res= arg->val_str(&tmp);
return (value_res ? (res ? sortcmp(value_res, res, cmp_charset) : 1) :
(res ? -1 : 0));
}
int compare(cmp_item *ci)
{
cmp_item_string *l_cmp= (cmp_item_string *) ci;
return sortcmp(value_res, l_cmp->value_res, cmp_charset);
}
cmp_item *make_same();
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
void set_charset(CHARSET_INFO *cs)
{
cmp_charset= cs;
value.set_quick(value_buff, sizeof(value_buff), cs);
}
2000-07-31 21:29:14 +02:00
};
class cmp_item_int :public cmp_item
{
longlong value;
public:
cmp_item_int() {} /* Remove gcc warning */
2000-07-31 21:29:14 +02:00
void store_value(Item *item)
{
value= item->val_int();
}
2000-07-31 21:29:14 +02:00
int cmp(Item *arg)
{
return value != arg->val_int();
}
int compare(cmp_item *ci)
{
cmp_item_int *l_cmp= (cmp_item_int *)ci;
return (value < l_cmp->value) ? -1 : ((value == l_cmp->value) ? 0 : 1);
}
cmp_item *make_same();
2000-07-31 21:29:14 +02:00
};
/*
Compare items in the DATETIME context.
Values are obtained with help of the get_datetime_value() function.
If the left item is a constant one then its value is cached in the
lval_cache variable.
*/
class cmp_item_datetime :public cmp_item
{
longlong value;
public:
THD *thd;
/* Item used for issuing warnings. */
Item *warn_item;
/* Cache for the left item. */
Item *lval_cache;
cmp_item_datetime(Item *warn_item_arg)
:thd(current_thd), warn_item(warn_item_arg), lval_cache(0) {}
void store_value(Item *item);
int cmp(Item *arg);
int compare(cmp_item *ci);
cmp_item *make_same();
};
2000-07-31 21:29:14 +02:00
class cmp_item_real :public cmp_item
{
double value;
public:
cmp_item_real() {} /* Remove gcc warning */
2000-07-31 21:29:14 +02:00
void store_value(Item *item)
{
2004-11-11 19:39:35 +01:00
value= item->val_real();
}
2000-07-31 21:29:14 +02:00
int cmp(Item *arg)
{
2004-11-11 19:39:35 +01:00
return value != arg->val_real();
}
int compare(cmp_item *ci)
{
cmp_item_real *l_cmp= (cmp_item_real *) ci;
return (value < l_cmp->value)? -1 : ((value == l_cmp->value) ? 0 : 1);
}
cmp_item *make_same();
2000-07-31 21:29:14 +02:00
};
2005-02-08 23:50:45 +01:00
class cmp_item_decimal :public cmp_item
{
my_decimal value;
public:
cmp_item_decimal() {} /* Remove gcc warning */
2005-02-08 23:50:45 +01:00
void store_value(Item *item);
int cmp(Item *arg);
int compare(cmp_item *c);
cmp_item *make_same();
};
/*
cmp_item for optimized IN with row (right part string, which never
be changed)
*/
class cmp_item_sort_string_in_static :public cmp_item_string
{
protected:
String value;
public:
cmp_item_sort_string_in_static(CHARSET_INFO *cs):
cmp_item_string(cs) {}
void store_value(Item *item)
{
value_res= item->val_str(&value);
}
int cmp(Item *item)
{
// Should never be called
DBUG_ASSERT(0);
return 1;
}
int compare(cmp_item *ci)
{
cmp_item_string *l_cmp= (cmp_item_string *) ci;
return sortcmp(value_res, l_cmp->value_res, cmp_charset);
}
cmp_item *make_same()
{
return new cmp_item_sort_string_in_static(cmp_charset);
}
};
2000-07-31 21:29:14 +02:00
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
/*
The class Item_func_case is the CASE ... WHEN ... THEN ... END function
implementation.
When there is no expression between CASE and the first WHEN
(the CASE expression) then this function simple checks all WHEN expressions
one after another. When some WHEN expression evaluated to TRUE then the
value of the corresponding THEN expression is returned.
When the CASE expression is specified then it is compared to each WHEN
expression individually. When an equal WHEN expression is found
corresponding THEN expression is returned.
In order to do correct comparisons several comparators are used. One for
each result type. Different result types that are used in particular
CASE ... END expression are collected in the fix_length_and_dec() member
function and only comparators for there result types are used.
*/
class Item_func_case :public Item_func
{
int first_expr_num, else_expr_num;
enum Item_result cached_result_type, left_result_type;
String tmp_value;
uint ncases;
Item_result cmp_type;
DTCollation cmp_collation;
enum_field_types cached_field_type;
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
cmp_item *cmp_items[5]; /* For all result types */
cmp_item *case_item;
public:
Item_func_case(List<Item> &list, Item *first_expr_arg, Item *else_expr_arg)
:Item_func(), first_expr_num(-1), else_expr_num(-1),
cached_result_type(INT_RESULT), left_result_type(INT_RESULT), case_item(0)
{
ncases= list.elements;
if (first_expr_arg)
{
first_expr_num= list.elements;
list.push_back(first_expr_arg);
}
if (else_expr_arg)
{
else_expr_num= list.elements;
list.push_back(else_expr_arg);
}
set_arguments(list);
bzero(&cmp_items, sizeof(cmp_items));
}
double val_real();
longlong val_int();
String *val_str(String *);
my_decimal *val_decimal(my_decimal *);
bool fix_fields(THD *thd, Item **ref);
void fix_length_and_dec();
uint decimal_precision() const;
table_map not_null_tables() const { return 0; }
enum Item_result result_type () const { return cached_result_type; }
enum_field_types field_type() const { return cached_field_type; }
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
const char *func_name() const { return "case"; }
virtual void print(String *str, enum_query_type query_type);
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
Item *find_item(String *str);
CHARSET_INFO *compare_collation() { return cmp_collation.collation; }
void cleanup();
void agg_str_lengths(Item *arg);
void agg_num_lengths(Item *arg);
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
};
/*
The Item_func_in class implements the in_expr IN(values_list) function.
The current implementation distinguishes 2 cases:
1) all items in the value_list are constants and have the same
result type. This case is handled by in_vector class.
2) items in the value_list have different result types or there is some
non-constant items.
In this case Item_func_in employs several cmp_item objects to performs
comparisons of in_expr and an item from the values_list. One cmp_item
object for each result type. Different result types are collected in the
fix_length_and_dec() member function by means of collect_cmp_types()
function.
*/
class Item_func_in :public Item_func_opt_neg
2000-07-31 21:29:14 +02:00
{
public:
/*
an array of values when the right hand arguments of IN
are all SQL constant and there are no nulls
*/
2000-07-31 21:29:14 +02:00
in_vector *array;
bool have_null;
/*
true when all arguments of the IN clause are of compatible types
and can be used safely as comparisons for key conditions
*/
bool arg_types_compatible;
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
Item_result left_result_type;
cmp_item *cmp_items[6]; /* One cmp_item for each result type */
DTCollation cmp_collation;
Item_func_in(List<Item> &list)
:Item_func_opt_neg(list), array(0), have_null(0),
arg_types_compatible(FALSE)
{
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
bzero(&cmp_items, sizeof(cmp_items));
allowed_arg_cols= 0; // Fetch this value from first argument
}
2000-07-31 21:29:14 +02:00
longlong val_int();
bool fix_fields(THD *, Item **);
2000-07-31 21:29:14 +02:00
void fix_length_and_dec();
uint decimal_precision() const { return 1; }
void cleanup()
{
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
uint i;
DBUG_ENTER("Item_func_in::cleanup");
Item_int_func::cleanup();
delete array;
array= 0;
for (i= 0; i <= (uint)DECIMAL_RESULT + 1; i++)
Fixed bug #18360: Type aggregation for IN and CASE may lead to a wrong result The IN function aggregates result types of all expressions. It uses that type in comparison of left expression and expressions in right part. This approach works in most cases. But let's consider the case when the right part contains both strings and integers. In that case this approach may cause wrong results because all strings which do not start with a digit are evaluated as 0. CASE uses the same approach when a CASE expression is given thus it's also affected. The idea behind this fix is to make IN function to compare expressions with different result types differently. For example a string in the left part will be compared as string with strings specified in right part and will be converted to real for comparison to int or real items in the right part. A new function called collect_cmp_types() is added. It collects different result types for comparison of first item in the provided list with each other item in the list. The Item_func_in class now can refer up to 5 cmp_item objects: 1 for each result type for comparison purposes. cmp_item objects are allocated according to found result types. The comparison of the left expression with any right part expression is now based only on result types of these expressions. The Item_func_case class is modified in the similar way when a CASE expression is specified. Now it can allocate up to 5 cmp_item objects to compare CASE expression with WHEN expressions of different types. The comparison of the CASE expression with any WHEN expression now based only on result types of these expressions.
2006-09-26 18:52:54 +02:00
{
delete cmp_items[i];
cmp_items[i]= 0;
}
DBUG_VOID_RETURN;
}
2000-07-31 21:29:14 +02:00
optimize_type select_optimize() const
{ return OPTIMIZE_KEY; }
virtual void print(String *str, enum_query_type query_type);
2000-07-31 21:29:14 +02:00
enum Functype functype() const { return IN_FUNC; }
const char *func_name() const { return " IN "; }
bool nulls_in_row();
2004-08-31 20:10:57 +02:00
bool is_bool_func() { return 1; }
CHARSET_INFO *compare_collation() { return cmp_collation.collation; }
2000-07-31 21:29:14 +02:00
};
class cmp_item_row :public cmp_item
{
cmp_item **comparators;
uint n;
public:
cmp_item_row(): comparators(0), n(0) {}
~cmp_item_row();
void store_value(Item *item);
2007-05-24 18:47:58 +02:00
inline void alloc_comparators();
int cmp(Item *arg);
int compare(cmp_item *arg);
cmp_item *make_same();
void store_value_by_template(cmp_item *tmpl, Item *);
2007-05-24 18:47:58 +02:00
friend void Item_func_in::fix_length_and_dec();
};
class in_row :public in_vector
{
cmp_item_row tmp;
public:
in_row(uint elements, Item *);
~in_row();
void set(uint pos,Item *item);
uchar *get_value(Item *item);
2007-05-24 18:47:58 +02:00
friend void Item_func_in::fix_length_and_dec();
Item_result result_type() { return ROW_RESULT; }
2000-07-31 21:29:14 +02:00
};
/* Functions used by where clause */
class Item_func_isnull :public Item_bool_func
{
protected:
2002-05-07 18:08:56 +02:00
longlong cached_value;
2000-07-31 21:29:14 +02:00
public:
Item_func_isnull(Item *a) :Item_bool_func(a) {}
longlong val_int();
enum Functype functype() const { return ISNULL_FUNC; }
void fix_length_and_dec()
{
decimals=0; max_length=1; maybe_null=0;
update_used_tables();
2000-07-31 21:29:14 +02:00
}
const char *func_name() const { return "isnull"; }
/* Optimize case of not_null_column IS NULL */
virtual void update_used_tables()
2000-07-31 21:29:14 +02:00
{
if (!args[0]->maybe_null)
{
used_tables_cache= 0; /* is always false */
const_item_cache= 1;
cached_value= (longlong) 0;
2000-07-31 21:29:14 +02:00
}
else
2002-05-07 18:08:56 +02:00
{
args[0]->update_used_tables();
if ((const_item_cache= !(used_tables_cache= args[0]->used_tables()) &&
!with_subselect))
{
/* Remember if the value is always NULL or never NULL */
cached_value= (longlong) args[0]->is_null();
}
2002-05-07 18:08:56 +02:00
}
2000-07-31 21:29:14 +02:00
}
table_map not_null_tables() const { return 0; }
2000-07-31 21:29:14 +02:00
optimize_type select_optimize() const { return OPTIMIZE_NULL; }
Item *neg_transformer(THD *thd);
CHARSET_INFO *compare_collation() { return args[0]->collation.collation; }
2000-07-31 21:29:14 +02:00
};
/* Functions used by HAVING for rewriting IN subquery */
class Item_in_subselect;
/*
This is like IS NOT NULL but it also remembers if it ever has
encountered a NULL.
*/
class Item_is_not_null_test :public Item_func_isnull
{
Item_in_subselect* owner;
public:
Item_is_not_null_test(Item_in_subselect* ow, Item *a)
:Item_func_isnull(a), owner(ow)
{}
enum Functype functype() const { return ISNOTNULLTEST_FUNC; }
longlong val_int();
const char *func_name() const { return "<is_not_null_test>"; }
void update_used_tables();
/*
we add RAND_TABLE_BIT to prevent moving this item from HAVING to WHERE
*/
table_map used_tables() const
{ return used_tables_cache | RAND_TABLE_BIT; }
};
2000-07-31 21:29:14 +02:00
class Item_func_isnotnull :public Item_bool_func
{
bool abort_on_null;
2000-07-31 21:29:14 +02:00
public:
Item_func_isnotnull(Item *a) :Item_bool_func(a), abort_on_null(0) {}
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return ISNOTNULL_FUNC; }
void fix_length_and_dec()
{
decimals=0; max_length=1; maybe_null=0;
}
2000-07-31 21:29:14 +02:00
const char *func_name() const { return "isnotnull"; }
optimize_type select_optimize() const { return OPTIMIZE_NULL; }
table_map not_null_tables() const
{ return abort_on_null ? not_null_tables_cache : 0; }
Item *neg_transformer(THD *thd);
virtual void print(String *str, enum_query_type query_type);
CHARSET_INFO *compare_collation() { return args[0]->collation.collation; }
void top_level_item() { abort_on_null=1; }
2000-07-31 21:29:14 +02:00
};
2000-07-31 21:29:14 +02:00
class Item_func_like :public Item_bool_func2
{
// Turbo Boyer-Moore data
bool canDoTurboBM; // pattern is '%abcd%' case
const char* pattern;
int pattern_len;
// TurboBM buffers, *this is owner
int* bmGs; // good suffix shift table, size is pattern_len + 1
int* bmBc; // bad character shift table, size is alphabet_size
void turboBM_compute_suffixes(int* suff);
void turboBM_compute_good_suffix_shifts(int* suff);
void turboBM_compute_bad_character_shifts();
bool turboBM_matches(const char* text, int text_len) const;
enum { alphabet_size = 256 };
Item *escape_item;
bool escape_used_in_parsing;
2000-07-31 21:29:14 +02:00
public:
int escape;
Item_func_like(Item *a,Item *b, Item *escape_arg, bool escape_used)
:Item_bool_func2(a,b), canDoTurboBM(FALSE), pattern(0), pattern_len(0),
bmGs(0), bmBc(0), escape_item(escape_arg),
escape_used_in_parsing(escape_used) {}
2000-07-31 21:29:14 +02:00
longlong val_int();
enum Functype functype() const { return LIKE_FUNC; }
optimize_type select_optimize() const;
cond_result eq_cmp_result() const { return COND_TRUE; }
const char *func_name() const { return "like"; }
bool fix_fields(THD *thd, Item **ref);
void cleanup();
2000-07-31 21:29:14 +02:00
};
class Item_func_regex :public Item_bool_func
{
my_regex_t preg;
2000-07-31 21:29:14 +02:00
bool regex_compiled;
bool regex_is_const;
String prev_regexp;
DTCollation cmp_collation;
CHARSET_INFO *regex_lib_charset;
int regex_lib_flags;
String conv;
int regcomp(bool send_error);
2000-07-31 21:29:14 +02:00
public:
Item_func_regex(Item *a,Item *b) :Item_bool_func(a,b),
regex_compiled(0),regex_is_const(0) {}
void cleanup();
2000-07-31 21:29:14 +02:00
longlong val_int();
bool fix_fields(THD *thd, Item **ref);
const char *func_name() const { return "regexp"; }
virtual inline void print(String *str, enum_query_type query_type)
{
print_op(str, query_type);
}
CHARSET_INFO *compare_collation() { return cmp_collation.collation; }
2000-07-31 21:29:14 +02:00
};
typedef class Item COND;
class Item_cond :public Item_bool_func
{
protected:
List<Item> list;
bool abort_on_null;
table_map and_tables_cache;
2000-07-31 21:29:14 +02:00
public:
/* Item_cond() is only used to create top level items */
Item_cond(): Item_bool_func(), abort_on_null(1)
{ const_item_cache=0; }
Item_cond(Item *i1,Item *i2)
:Item_bool_func(), abort_on_null(0)
{
list.push_back(i1);
list.push_back(i2);
}
Item_cond(THD *thd, Item_cond *item);
Item_cond(List<Item> &nlist)
:Item_bool_func(), list(nlist), abort_on_null(0) {}
bool add(Item *item)
{
DBUG_ASSERT(item);
return list.push_back(item);
}
bool add_at_head(Item *item)
{
DBUG_ASSERT(item);
return list.push_front(item);
}
void add_at_head(List<Item> *nlist)
{
DBUG_ASSERT(nlist->elements);
list.prepand(nlist);
}
bool fix_fields(THD *, Item **ref);
2000-07-31 21:29:14 +02:00
enum Type type() const { return COND_ITEM; }
List<Item>* argument_list() { return &list; }
table_map used_tables() const;
void update_used_tables();
virtual void print(String *str, enum_query_type query_type);
void split_sum_func(THD *thd, Item **ref_pointer_array, List<Item> &fields);
friend int setup_conds(THD *thd, TABLE_LIST *tables, TABLE_LIST *leaves,
COND **conds);
void top_level_item() { abort_on_null=1; }
void copy_andor_arguments(THD *thd, Item_cond *item);
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
bool walk(Item_processor processor, bool walk_subquery, uchar *arg);
Item *transform(Item_transformer transformer, uchar *arg);
void traverse_cond(Cond_traverser, void *arg, traverse_order order);
void neg_arguments(THD *thd);
enum_field_types field_type() const { return MYSQL_TYPE_LONGLONG; }
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
bool subst_argument_checker(uchar **arg) { return TRUE; }
Item *compile(Item_analyzer analyzer, uchar **arg_p,
Item_transformer transformer, uchar *arg_t);
2000-07-31 21:29:14 +02:00
};
/*
2005-02-08 23:50:45 +01:00
The class Item_equal is used to represent conjunctions of equality
predicates of the form field1 = field2, and field=const in where
conditions and on expressions.
All equality predicates of the form field1=field2 contained in a
2005-02-08 23:50:45 +01:00
conjunction are substituted for a sequence of items of this class.
An item of this class Item_equal(f1,f2,...fk) represents a
multiple equality f1=f2=...=fk.
2005-02-08 23:50:45 +01:00
If a conjunction contains predicates f1=f2 and f2=f3, a new item of
this class is created Item_equal(f1,f2,f3) representing the multiple
equality f1=f2=f3 that substitutes the above equality predicates in
2005-02-08 23:50:45 +01:00
the conjunction.
A conjunction of the predicates f2=f1 and f3=f1 and f3=f2 will be
substituted for the item representing the same multiple equality
f1=f2=f3.
2005-02-08 23:50:45 +01:00
An item Item_equal(f1,f2) can appear instead of a conjunction of
f2=f1 and f1=f2, or instead of just the predicate f1=f2.
2005-02-08 23:50:45 +01:00
An item of the class Item_equal inherits equalities from outer
conjunctive levels.
Suppose we have a where condition of the following form:
WHERE f1=f2 AND f3=f4 AND f3=f5 AND ... AND (...OR (f1=f3 AND ...)).
In this case:
f1=f2 will be substituted for Item_equal(f1,f2);
f3=f4 and f3=f5 will be substituted for Item_equal(f3,f4,f5);
f1=f3 will be substituted for Item_equal(f1,f2,f3,f4,f5);
An object of the class Item_equal can contain an optional constant
2005-02-08 23:50:45 +01:00
item c. Then it represents a multiple equality of the form
c=f1=...=fk.
Objects of the class Item_equal are used for the following:
1. An object Item_equal(t1.f1,...,tk.fk) allows us to consider any
pair of tables ti and tj as joined by an equi-condition.
Thus it provide us with additional access paths from table to table.
2. An object Item_equal(t1.f1,...,tk.fk) is applied to deduce new
SARGable predicates:
f1=...=fk AND P(fi) => f1=...=fk AND P(fi) AND P(fj).
It also can give us additional index scans and can allow us to
improve selectivity estimates.
3. An object Item_equal(t1.f1,...,tk.fk) is used to optimize the
selected execution plan for the query: if table ti is accessed
before the table tj then in any predicate P in the where condition
2005-02-08 23:50:45 +01:00
the occurrence of tj.fj is substituted for ti.fi. This can allow
an evaluation of the predicate at an earlier step.
When feature 1 is supported they say that join transitive closure
is employed.
When feature 2 is supported they say that search argument transitive
closure is employed.
Both features are usually supported by preprocessing original query and
adding additional predicates.
We do not just add predicates, we rather dynamically replace some
predicates that can not be used to access tables in the investigated
plan for those, obtained by substitution of some fields for equal fields,
that can be used.
Prepared Statements/Stored Procedures note: instances of class
Item_equal are created only at the time a PS/SP is executed and
are deleted in the end of execution. All changes made to these
objects need not be registered in the list of changes of the parse
tree and do not harm PS/SP re-execution.
Item equal objects are employed only at the optimize phase. Usually they are
not supposed to be evaluated. Yet in some cases we call the method val_int()
for them. We have to take care of restricting the predicate such an
object represents f1=f2= ...=fn to the projection of known fields fi1=...=fik.
*/
class Item_equal: public Item_bool_func
{
List<Item_field> fields; /* list of equal field items */
Item *const_item; /* optional constant item equal to fields items */
cmp_item *eval_item;
Arg_comparator cmp;
bool cond_false;
bool compare_as_dates;
public:
inline Item_equal()
: Item_bool_func(), const_item(0), eval_item(0), cond_false(0)
{ const_item_cache=0 ;}
Item_equal(Item_field *f1, Item_field *f2);
Item_equal(Item *c, Item_field *f);
Item_equal(Item_equal *item_equal);
inline Item* get_const() { return const_item; }
void compare_const(Item *c);
void add(Item *c, Item_field *f);
void add(Item *c);
void add(Item_field *f);
uint members();
bool contains(Field *field);
Item_field* get_first() { return fields.head(); }
void merge(Item_equal *item);
void update_const();
enum Functype functype() const { return MULT_EQUAL_FUNC; }
longlong val_int();
const char *func_name() const { return "multiple equal"; }
optimize_type select_optimize() const { return OPTIMIZE_EQUAL; }
void sort(Item_field_cmpfunc compare, void *arg);
friend class Item_equal_iterator;
void fix_length_and_dec();
bool fix_fields(THD *thd, Item **ref);
void update_used_tables();
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
bool walk(Item_processor processor, bool walk_subquery, uchar *arg);
Item *transform(Item_transformer transformer, uchar *arg);
virtual void print(String *str, enum_query_type query_type);
CHARSET_INFO *compare_collation()
{ return fields.head()->collation.collation; }
};
class COND_EQUAL: public Sql_alloc
{
public:
uint max_members; /* max number of members the current level
list and all lower level lists */
COND_EQUAL *upper_levels; /* multiple equalities of upper and levels */
List<Item_equal> current_level; /* list of multiple equalities of
the current and level */
COND_EQUAL()
{
upper_levels= 0;
}
};
2004-11-29 22:47:50 +01:00
class Item_equal_iterator : public List_iterator_fast<Item_field>
{
public:
inline Item_equal_iterator(Item_equal &item_equal)
:List_iterator_fast<Item_field> (item_equal.fields)
{}
inline Item_field* operator++(int)
{
Item_field *item= (*(List_iterator_fast<Item_field> *) this)++;
return item;
}
inline void rewind(void)
{
List_iterator_fast<Item_field>::rewind();
}
};
2000-07-31 21:29:14 +02:00
class Item_cond_and :public Item_cond
{
public:
COND_EQUAL cond_equal; /* contains list of Item_equal objects for
the current and level and reference
to multiple equalities of upper and levels */
2000-07-31 21:29:14 +02:00
Item_cond_and() :Item_cond() {}
Item_cond_and(Item *i1,Item *i2) :Item_cond(i1,i2) {}
Item_cond_and(THD *thd, Item_cond_and *item) :Item_cond(thd, item) {}
Item_cond_and(List<Item> &list_arg): Item_cond(list_arg) {}
2000-07-31 21:29:14 +02:00
enum Functype functype() const { return COND_AND_FUNC; }
longlong val_int();
const char *func_name() const { return "and"; }
table_map not_null_tables() const
{ return abort_on_null ? not_null_tables_cache: and_tables_cache; }
Item* copy_andor_structure(THD *thd)
{
Item_cond_and *item;
2004-02-05 18:14:48 +01:00
if ((item= new Item_cond_and(thd, this)))
item->copy_andor_arguments(thd, this);
return item;
}
Item *neg_transformer(THD *thd);
2000-07-31 21:29:14 +02:00
};
Bug#30237 (Performance regression in boolean expressions) This is a performance bug, related to the parsing or 'OR' and 'AND' boolean expressions. Let N be the number of expressions involved in a OR (respectively AND). When N=1 For example, "select 1" involve only 1 term: there is no OR operator. In 4.0 and 4.1, parsing expressions not involving OR had no overhead. In 5.0, parsing adds some overhead, with Select->expr_list. With this patch, the overhead introduced in 5.0 has been removed, so that performances for N=1 should be identical to the 4.0 performances, which are optimal (there is no code executed at all) The overhead in 5.0 was in fact affecting significantly some operations. For example, loading 1 Million rows into a table with INSERTs, for a table that has 100 columns, leads to parsing 100 Millions of expressions, which means that the overhead related to Select->expr_list is executed 100 Million times ... Considering that N=1 is by far the most probable expression, this case should be optimal. When N=2 For example, "select a OR b" involves 2 terms in the OR operator. In 4.0 and 4.1, parsing expressions involving 2 terms created 1 Item_cond_or node, which is the expected result. In 5.0, parsing these expression also produced 1 node, but with some extra overhead related to Select->expr_list : creating 1 list in Select->expr_list and another in Item_cond::list is inefficient. With this patch, the overhead introduced in 5.0 has been removed so that performances for N=2 should be identical to the 4.0 performances. Note that the memory allocation uses the new (thd->mem_root) syntax directly. The cost of "is_cond_or" is estimated to be neglectable: the real problem of the performance degradation comes from unneeded memory allocations. When N>=3 For example, "select a OR b OR c ...", which involves 3 or more terms. In 4.0 and 4.1, the parser had no significant cost overhead, but produced an Item tree which is difficult to evaluate / optimize during runtime. In 5.0, the parser produces a better Item tree, using the Item_cond constructor that accepts a list of children directly, but at an extra cost related to Select->expr_list. With this patch, the code is implemented to take the best of the two implementations: - there is no overhead with Select->expr_list - the Item tree generated is optimized and flattened. This is achieved by adding children nodes into the Item tree directly, with Item_cond::add(), which avoids the need for temporary lists and memory allocation Note that this patch also provide an extra optimization, that the previous code in 5.0 did not provide: expressions are flattened in the Item tree, based on what the expression already parsed is, and not based on the order in which rules are reduced. For example : "(a OR b) OR c", "a OR (b OR c)" would both be represented with 2 Item_cond_or nodes before this patch, and with 1 node only with this patch. The logic used is based on the mathematical properties of the OR operator (it's associative), and produces a simpler tree.
2007-08-22 19:05:35 +02:00
inline bool is_cond_and(Item *item)
{
if (item->type() != Item::COND_ITEM)
return FALSE;
Item_cond *cond_item= (Item_cond*) item;
return (cond_item->functype() == Item_func::COND_AND_FUNC);
}
2000-07-31 21:29:14 +02:00
class Item_cond_or :public Item_cond
{
public:
Item_cond_or() :Item_cond() {}
Item_cond_or(Item *i1,Item *i2) :Item_cond(i1,i2) {}
Item_cond_or(THD *thd, Item_cond_or *item) :Item_cond(thd, item) {}
Item_cond_or(List<Item> &list_arg): Item_cond(list_arg) {}
2000-07-31 21:29:14 +02:00
enum Functype functype() const { return COND_OR_FUNC; }
longlong val_int();
const char *func_name() const { return "or"; }
table_map not_null_tables() const { return and_tables_cache; }
Item* copy_andor_structure(THD *thd)
{
Item_cond_or *item;
2004-02-05 18:14:48 +01:00
if ((item= new Item_cond_or(thd, this)))
item->copy_andor_arguments(thd, this);
return item;
}
Item *neg_transformer(THD *thd);
2000-07-31 21:29:14 +02:00
};
Bug#30237 (Performance regression in boolean expressions) This is a performance bug, related to the parsing or 'OR' and 'AND' boolean expressions. Let N be the number of expressions involved in a OR (respectively AND). When N=1 For example, "select 1" involve only 1 term: there is no OR operator. In 4.0 and 4.1, parsing expressions not involving OR had no overhead. In 5.0, parsing adds some overhead, with Select->expr_list. With this patch, the overhead introduced in 5.0 has been removed, so that performances for N=1 should be identical to the 4.0 performances, which are optimal (there is no code executed at all) The overhead in 5.0 was in fact affecting significantly some operations. For example, loading 1 Million rows into a table with INSERTs, for a table that has 100 columns, leads to parsing 100 Millions of expressions, which means that the overhead related to Select->expr_list is executed 100 Million times ... Considering that N=1 is by far the most probable expression, this case should be optimal. When N=2 For example, "select a OR b" involves 2 terms in the OR operator. In 4.0 and 4.1, parsing expressions involving 2 terms created 1 Item_cond_or node, which is the expected result. In 5.0, parsing these expression also produced 1 node, but with some extra overhead related to Select->expr_list : creating 1 list in Select->expr_list and another in Item_cond::list is inefficient. With this patch, the overhead introduced in 5.0 has been removed so that performances for N=2 should be identical to the 4.0 performances. Note that the memory allocation uses the new (thd->mem_root) syntax directly. The cost of "is_cond_or" is estimated to be neglectable: the real problem of the performance degradation comes from unneeded memory allocations. When N>=3 For example, "select a OR b OR c ...", which involves 3 or more terms. In 4.0 and 4.1, the parser had no significant cost overhead, but produced an Item tree which is difficult to evaluate / optimize during runtime. In 5.0, the parser produces a better Item tree, using the Item_cond constructor that accepts a list of children directly, but at an extra cost related to Select->expr_list. With this patch, the code is implemented to take the best of the two implementations: - there is no overhead with Select->expr_list - the Item tree generated is optimized and flattened. This is achieved by adding children nodes into the Item tree directly, with Item_cond::add(), which avoids the need for temporary lists and memory allocation Note that this patch also provide an extra optimization, that the previous code in 5.0 did not provide: expressions are flattened in the Item tree, based on what the expression already parsed is, and not based on the order in which rules are reduced. For example : "(a OR b) OR c", "a OR (b OR c)" would both be represented with 2 Item_cond_or nodes before this patch, and with 1 node only with this patch. The logic used is based on the mathematical properties of the OR operator (it's associative), and produces a simpler tree.
2007-08-22 19:05:35 +02:00
inline bool is_cond_or(Item *item)
{
if (item->type() != Item::COND_ITEM)
return FALSE;
Item_cond *cond_item= (Item_cond*) item;
return (cond_item->functype() == Item_func::COND_OR_FUNC);
}
2000-07-31 21:29:14 +02:00
/*
2005-02-08 23:50:45 +01:00
XOR is Item_cond, not an Item_int_func because we could like to
optimize (a XOR b) later on. It's low prio, though
*/
2003-08-11 21:44:43 +02:00
class Item_cond_xor :public Item_cond
{
public:
Item_cond_xor() :Item_cond() {}
Item_cond_xor(Item *i1,Item *i2) :Item_cond(i1,i2) {}
enum Functype functype() const { return COND_XOR_FUNC; }
/* TODO: remove the next line when implementing XOR optimization */
enum Type type() const { return FUNC_ITEM; }
longlong val_int();
const char *func_name() const { return "xor"; }
void top_level_item() {}
};
2005-02-08 23:50:45 +01:00
/* Some useful inline functions */
2000-07-31 21:29:14 +02:00
inline Item *and_conds(Item *a, Item *b)
2000-07-31 21:29:14 +02:00
{
if (!b) return a;
if (!a) return b;
return new Item_cond_and(a, b);
2000-07-31 21:29:14 +02:00
}
2002-11-21 14:56:48 +01:00
Item *and_expressions(Item *a, Item *b, Item **org_item);
longlong get_datetime_value(THD *thd, Item ***item_arg, Item **cache_arg,
Item *warn_item, bool *is_null);
2009-11-06 15:20:27 +01:00
bool get_mysql_time_from_str(THD *thd, String *str, timestamp_type warn_type,
const char *warn_name, MYSQL_TIME *l_time);
/*
These need definitions from this file but the variables are defined
in mysqld.h. The variables really belong in this component, but for
the time being we leave them in mysqld.cc to avoid merge problems.
*/
extern Eq_creator eq_creator;
extern Ne_creator ne_creator;
extern Gt_creator gt_creator;
extern Lt_creator lt_creator;
extern Ge_creator ge_creator;
extern Le_creator le_creator;
#endif /* ITEM_CMPFUNC_INCLUDED */