With INTERSECT/EXCEPT fact that subquery item of IN/ALL/ANY was not assigned value does not mean that temporary table used for calculating unit is empty (records could be deleted).
As a result of this merge the code for the following tasks appears in 10.3:
- MDEV-12172 Implement tables specified by table value constructors
- MDEV-12176 Transform [NOT] IN predicate with long list of values INTO
[NOT] IN subquery.
- Added sql/mariadb.h file that should be included first by files in sql
directory, if sql_plugin.h is not used (sql_plugin.h adds SHOW variables
that must be done before my_global.h is included)
- Removed a lot of include my_global.h from include files
- Removed include's of some files that my_global.h automatically includes
- Removed duplicated include's of my_sys.h
- Replaced include my_config.h with my_global.h
"Optimization for equi-joins of derived tables with GROUP BY"
should be considered rather as a 'proof of concept'.
The task itself is targeted at an optimization that employs re-writing
equi-joins with grouping derived tables / views into lateral
derived tables. Here's an example of such transformation:
select t1.a,t.max,t.min
from t1 [left] join
(select a, max(t2.b) max, min(t2.b) min from t2
group by t2.a) as t
on t1.a=t.a;
=>
select t1.a,tl.max,tl.min
from t1 [left] join
lateral (select a, max(t2.b) max, min(t2.b) min from t2
where t1.a=t2.a) as t
on 1=1;
The transformation pushes the equi-join condition t1.a=t.a into the
derived table making it dependent on table t1. It means that for
every row from t1 a new derived table must be filled out. However
the size of any of these derived tables is just a fraction of the
original derived table t. One could say that transformation 'splits'
the rows used for the GROUP BY operation into separate groups
performing aggregation for a group only in the case when there is
a match for the current row of t1.
Apparently the transformation may produce a query with a better
performance only in the case when
- the GROUP BY list refers only to fields returned by the derived table
- there is an index I on one of the tables T used in FROM list of
the specification of the derived table whose prefix covers the
the fields from the proper beginning of the GROUP BY list or
fields that are equal to those fields.
Whether the result of the re-writing can be executed faster depends
on many factors:
- the size of the original derived table
- the size of the table T
- whether the index I is clustering for table T
- whether the index I fully covers the GROUP BY list.
This patch only tries to improve the chosen execution plan using
this transformation. It tries to do it only when the chosen
plan reaches the derived table by a key whose prefix covers
all the fields of the derived table produced by the fields of
the table T from the GROUP BY list.
The code of the patch does not evaluates the cost of the improved
plan. If certain conditions are met the transformation is applied.
TVC can be used in UNION-statement, in view and in subquery.
Files where TVC is defined and its methods are stored added.
Methods exec and prepare for TVC added.
Tests for TVC added.
prepare of "fake_select" for union made in JOIN::prepare only if
we do not execute it then before reset, i.e it was for PS prepare
and now required for CREATE VIEW to make global ORDER BY which
belongs to "fake_select" prepared.
GROUP BY
Issue 1:
--------
This problem occurs in the following conditions:
1) A UNION is present in the subquery of select list and
handles multiple columns.
2) Query has a GROUP BY.
A temporary table is created to handle the UNION.
Item_field objects are based on the expressions of the
result of the UNION (ie. the fake_select_lex). While
checking validity of the columns in the GROUP BY list, the
columns of the temporary table are checked in
Item_ident::local_column. But the Item_field objects
created for the temporary table don't have information like
the Name_resolution_context that they belong to or whether
they are dependent on an outer query. Since these members
are null, incorrect behavior is caused.
This can happen when such Item objects are cached to apply
the IN-to-EXISTS transform for Item_row.
Solution to Issue 1:
--------------------
Context information of the first select in the UNION will
be assigned to the new Item_field objects.
Issue 2:
--------
This problem occurs in the following conditions:
1) A UNION is present in the subquery of select list.
2) A column in the UNION's first SELECT refers to a table
in the outer-query making it a dependent union.
3) GROUP BY column refers to the outer-referencing column.
While resolving the select list with an outer-reference, an
Item_outer_ref object is created to handle the
outer-query's GROUP BY list. The Item_outer_ref object
replaces the Item_field object in the item tree.
Item_outer_ref::fix_fields will be called only while fixing
the inner references of the outer query.
Before resolving the outer-query, an Item_type_holder
object needs to be created to handle the UNION. But as
explained above, the Item_outer_ref object has not been
fixed yet. Having a fixed Item object is a pre-condition
for creating an Item_type_holder.
Solution to Issue 2:
--------------------
Use the reference (real_item()) of an Item_outer_ref object
instead of the object itself while creating an
Item_type_holder.
Introducing a new class Type_holder (used internally in sql_union.cc),
to reuse exactly the same data type attribute aggregation Type_handler API
for hybrid functions and UNION.
This fixes a number of bugs in UNION:
- MDEV-9495 Wrong field type for a UNION of a signed and an unsigned INT expression
- MDEV-9497 UNION and COALESCE produce different field types for DECIMAL+INT
- MDEV-12594 UNION between fixed length double columns does not always preserve scale
- MDEV-12595 UNION converts INT to BIGINT
- MDEV-12599 UNION is not symmetric when mixing INT and CHAR
Details:
- sql_union.cc: Reusing attribute aggregation for UNION.
Adding new methods:
* st_select_lex_unit::join_union_type_handlers()
* st_select_lex_unit::join_union_type_attributes()
* st_select_lex_unit::join_union_item_types()
Removing the old join_types()-based code.
- Changing Type_handler::Item_hybrid_func_fix_attributes()
to accept "name", Type_handler_hybrid_field_type, Type_all_attributes
as three separate parameters instead of a single Item_hybrid_func parameter,
to make it possible to pass both Item_hybrid_func and Type_holder.
- Moving the former special GEOMETRY and ENUM/SET attribute aggregation code
from Item_type_holder::join_types() to
* Type_handler_typelib::Item_hybrid_func_fix_attributes().
* Type_handler_geometry::Item_hybrid_func_fix_attrubutes().
This makes GEOMETRY/ENUM/SET symmetric with all other data types
(from the UNION point of view).
Removing Item_type_holder::join_types() and Item_type_holder::get_full_info().
- Adding new methods into Type_all_attributes:
* Type_all_attributes::set_geometry_type() and
Item_hybrid_func::set_geometry_type().
* Adding Type_all_attributes::get_typelib().
* Adding Type_all_attributes::set_typelib().
- Adding Type_handler_typelib as a common parent for
Type_handler_enum and Type_handler_set, to avoid code duplication: they have
already had two common methods, and we're adding one more shared method.
- Adding Type_all_attributes::set_maybe_null(), as some type handlers
may want to set maybe_null (e.g. Type_handler_geometry) during data type
attribute aggregation.
- Changing Type_geometry_attributes() to accept Type_handler
and Type_all_attributes as two separate parameters, instead
of a single Item parameter, to make it possible to pass Type_holder.
- Adding Item_args::add_argument().
- Moving Item_args::alloc_arguments() from "protected" to "public".
- Moving Item_type_holder::Item_type_holder() from item.cc to item.h, as
now it's very simple.
Btw, this constructor should probably be eventually removed.
It's now used only in sql_show.cc, which could be modified to use
Item_return_decimal (for symmetry with Item_return_xxx created for all
other data types). Or, another option: remove all Item_return_xxx and
use Item_type_holder for all data types instead.
- storage/tokudb/mysql-test/tokudb/r/type_float.result
Recording new results (MDEV-12594).
- mysql-test/r/cte_recursive.result
Recording new results (MDEV-9497)
- mysql-test/r/subselect*.result
Recording new results (MDEV-12595)
- mysql-test/r/metadata.result
Recording new results (MDEV-9495)
- mysql-test/r/temp_table.result
Recording new results (MDEV-12594)
- mysql-test/r/type_float.result
Recording new results (MDEV-12594)
When the rows produced on the current iteration are sent to the
temporary table T of the UNION type created for CTE the rows
that were not there simultaneously are sent to the temporary
table D that contains rows for the next iteration. The test
whether a row was in T checks the return code of writing into T.
If just a HEAP table is used for T then the return code is
HA_ERR_FOUND_DUPP_KEY, but if an ARIA table is used for T then
the return code is HA_ERR_FOUND_DUPP_UNIQUE.
The implementation of select_union_recursive::send_data()
erroneously checked only for the first return code. So if an Aria
table was used for T then all rows produced by the current iteration
went to D and and in most cases D grew with each iteration.
Whether T has reached stabilization is detected by
checking whether D is empty. So as a result, the iterations were
never stopped unless a limit for them was set.
Fixed by checking for both HA_ERR_FOUND_DUPP_KEY and
HA_ERR_FOUND_DUPP_UNIQUE as return codes returned by
the function writing a row into the temporary table T.
Benefits of this patch:
- Removed a lot of calls to strlen(), especially for field_string
- Strings generated by parser are now const strings, less chance of
accidently changing a string
- Removed a lot of calls with LEX_STRING as parameter (changed to pointer)
- More uniform code
- Item::name_length was not kept up to date. Now fixed
- Several bugs found and fixed (Access to null pointers,
access of freed memory, wrong arguments to printf like functions)
- Removed a lot of casts from (const char*) to (char*)
Changes:
- This caused some ABI changes
- lex_string_set now uses LEX_CSTRING
- Some fucntions are now taking const char* instead of char*
- Create_field::change and after changed to LEX_CSTRING
- handler::connect_string, comment and engine_name() changed to LEX_CSTRING
- Checked printf() related calls to find bugs. Found and fixed several
errors in old code.
- A lot of changes from LEX_STRING to LEX_CSTRING, especially related to
parsing and events.
- Some changes from LEX_STRING and LEX_STRING & to LEX_CSTRING*
- Some changes for char* to const char*
- Added printf argument checking for my_snprintf()
- Introduced null_clex_str, star_clex_string, temp_lex_str to simplify
code
- Added item_empty_name and item_used_name to be able to distingush between
items that was given an empty name and items that was not given a name
This is used in sql_yacc.yy to know when to give an item a name.
- select table_name."*' is not anymore same as table_name.*
- removed not used function Item::rename()
- Added comparision of item->name_length before some calls to
my_strcasecmp() to speed up comparison
- Moved Item_sp_variable::make_field() from item.h to item.cc
- Some minimal code changes to avoid copying to const char *
- Fixed wrong error message in wsrep_mysql_parse()
- Fixed wrong code in find_field_in_natural_join() where real_item() was
set when it shouldn't
- ER_ERROR_ON_RENAME was used with extra arguments.
- Removed some (wrong) ER_OUTOFMEMORY, as alloc_root will already
give the error.
TODO:
- Check possible unsafe casts in plugin/auth_examples/qa_auth_interface.c
- Change code to not modify LEX_CSTRING for database name
(as part of lower_case_table_names)
This patch fixed some problems that occurred with subqueries that
contained directly or indirectly recursive references to recursive CTEs.
1. A [NOT] IN predicate with a constant left operand and a non-correlated
subquery as the right operand used in the specification of a recursive CTE
was considered as a constant predicate and was evaluated only once.
Now such a predicate is re-evaluated after every iteration of the process
that produces the records of the recursive CTE.
2. The Exists-To-IN transformation could be applied to [NOT] IN predicates
with recursive references. This opened a possibility of materialization
for the subqueries used as right operands. Yet, materialization
is prohibited for the subqueries if they contain a recursive reference.
Now the Exists-To-IN transformation cannot be applied for subquery
predicates with recursive references.
The function st_select_lex::check_subqueries_with_recursive_references()
is called now only for the first execution of the SELECT.
The function st_select_lex_unit::exec_recursive() incorrectly determined
that a CTE mutually recursive with some others was stabilized in the case
when the non-recursive part of the CTE returned an empty set. As a result
the server fell into an infinite loop when executing a query using
this CTE.
The temporary tables created for recursive table references
should be closed in close_thread_tables(), because they might
be used in the statements like ANALYZE WITH r AS (...) SELECT * from r
where r is defined through recursion.
1. The rows of a recursive CTE at some point may overflow
the HEAP temporary table containing them. At this point
the table is converted to a MyISAM temporary table and the
new added rows are placed into this MyISAM table.
A bug in the of select_union_recursive::send_data prevented
the server from writing the row that caused the overflow
into the temporary table used for the result of the iteration
steps. This could lead, in particular,to a premature end
of the iterations.
2. The method TABLE::insert_all_rows_into() that was used
to copy all rows of one temporary table into another
did not take into account that the destination temporary
table must be converted to a MyISAM table at some point.
This patch fixed this problem. It also renamed the method
into TABLE::insert_all_rows_into_tmp_table() and added
an extra parameter needed for the conversion.
The idea of this fix was taken from the patch by Roy Lyseng
for mysql-5.6 bug iBug#14740889: "Wrong result for aggregate
functions when executing query through cursor".
Here's Roy's comment for his patch:
"
The problem was that a grouped query did not behave properly when
executed using a cursor. On further inspection, the query used one
intermediate temporary table for the grouping.
Then, Select_materialize::send_result_set_metadata created a temporary
table for storing the query result. Notice that get_unit_column_types()
is used to retrieve column meta-data for the query. The items contained
in this list are later modified so that their result_field points to
the row buffer of the materialized temporary table for the cursor.
But prior to this, these result_field objects have been prepared for
use in the grouping operation, by JOIN::make_tmp_tables_info(), hence
the grouping operation operates on wrong column buffers.
The problem is solved by using the list JOIN::fields when copying data
to the materialized table. This list is set by JOIN::make_tmp_tables_info()
and points to the columns of the last intermediate temporary table of
the executed query. For a UNION, it points to the temporary table
that is the result of the UNION query.
Notice that we have to assign a value to ::fields early in JOIN::optimize()
in case the optimization shortcuts due to a const plan detection.
A more optimal solution might be to avoid creating the final temporary
table when the query result is already stored in a temporary table.
"
The patch does not contain a test case, but the description of the
problem corresponds exactly what could be observed in the test
case for mdev-11081.
Added comments.
Added reaction for exceeding maximum number of elements in with clause.
Added a test case to check this reaction.
Added a test case where the specification of a recursive table
uses two non-recursive with tables.
Moved checking whether the limit set for the number of iterations
when executing a recursive query has been reached from
st_select_lex_unit::exec_recursive to TABLE_LIST::fill_recursive.
Changed the name of the system variable max_recursion_level for
max_recursive_iterations.
Adjusted test cases.
- Tabular EXPLAIN now prints "RECURSIVE UNION".
- There is a basic implementation of EXPLAIN FORMAT=JSON.
- it produces "recursive_union" JSON struct
- No other details or ANALYZE support, yet.
Temporary tables created for recursive CTE
were instantiated at the prepare phase. As
a result these temporary tables missed
indexes for look-ups and optimizer could not
use them.
Actually mutually recursive CTE were not functional. Now the code
for mutually recursive CTE looks like functional, but still needs
re-writing.
Added many new test cases for mutually recursive CTE.
Added test cases to check the fix.
Fixed the problem of wrong types of recursive tables when the type of anchor part does not coincide with the
type of recursive part.
Prevented usage of marerialization and subquery cache for subqueries with recursive references.
Introduced system variables 'max_recursion_level'.
Added a test case to test usage of this variable.
