Analysis:
The cause of the valgrind warning was an attempt to evaluate a Field that was not yet read.
The reason was that on one hand Item_func_isnotnull was marked as constant by
Item_func_isnotnull::update_used_tables, and this allowed eval_const_cond() to be called.
On the other hand Item_func_isnotnull::val_int() evaluated its argument as if it was not
constant.
Solution:
The fix make sure that Item_func_isnotnull::val_int() doesn't evaluate its argument when
it is constant and cannot be NULL, because the result is known in this case.
This patch almost totally revised the patch for bug mdev-4177.
The latter had too many defects. In particular, it did not
propagate multiple equalities formed when merging a degenerate
disjunct into underlying AND formula.
- Modify the way Item_cond::fix_fields() and Item_cond::eval_not_null_tables()
calculate bitmap for Item_cond_or::not_null_tables():
if they see a "... OR inexpensive_const_false_item OR ..." then the item can
be ignored.
- Updated test results. There can be more warnings produced since parts of WHERE
are evaluated more times.
includes:
* remove some remnants of "Bug#14521864: MYSQL 5.1 TO 5.5 BUGS PARTITIONING"
* introduce LOCK_share, now LOCK_ha_data is strictly for engines
* rea_create_table() always creates .par file (even in "frm-only" mode)
* fix a 5.6 bug, temp file leak on dummy ALTER TABLE
Cleanup: remove TIME_FUZZY_DATE.
Introduce TIME_FUZZY_DATES which means "very fuzzy, the resulting
value is only used for comparison. It can be invalid date, fine, as long as it can be
compared".
Updated many tests results (they're better now).
get_datetime_value() should not double-cache its own Item_cache_temporal items,
but it *should* cache other Item_cache items, such as Item_cache_str.
sql/item.h:
shortcut, to avoid going through the switch in Item::cmp_type()
sql/item_cmpfunc.cc:
even if the item is Item_cache_str - it still needs to be converted and cached.
sql/item_timefunc.h:
all descendants of Item_temporal_func always have cmp_type==TIME_RESULT.
Even Item_date_add_interval, that might have field_type == MYSQL_TYPE_STRING.
IN IN-CLAUSE USING MYISAM OR MEMORY ENGINE
Backport from 5.6. Original message:
The coincidences caused a data loss:
* The query has IN subqueries nested twice,
* the WHERE clause of the inner subquery refers to the
outer field, and the whole WHERE clause returns FALSE,
* the inner subquery has a LEFT JOIN that joins a single
row with a row of NULLs; one of that NULL columns
represents the select list of the subquery.
Normally, that inner subquery should return empty record set.
However, in our case:
* the Item_is_not_null_test item goes constant, since
its underlying field is NULL (because of LEFT JOIN ... ON
FALSE of const table row with a row of nulls);
* we evaluate Item_is_not_null_test::val_int() as a part
of fake HAVING expression of the transformed subquery;
* as far as the underlying field is NULL, we optimize
out the whole fake HAVING expression as FALSE as well
as a whole subquery with a zero result:
Impossible HAVING noticed after reading const tables";
* thus, the optimizer ignores the presence of the WHERE
clause (the WHERE expression is FALSE in our case, so
the subquery should return empty set);
* however, during the evaluation of the
Item_is_not_null_test::val_int() in the optimizer,
it marked its "owner" with the "was_null" flag -- that
forced the subquery to return UNKNOWN instead of empty
set.
That caused a wrong result.
The problem is a regression of the small cleanup in
the fix for the bug11827369 (the Item_is_not_null_test part)
that conflicts with optimizations in the fix for the bug11752543.
Before that regression the Item_is_not_null_test items
never were constants.
The fix is the rollback of Item_is_not_null_test parts
of the bug11827369 fix.
The function remove_eq_cond removes the parts of a disjunction
for which it has been proved that they are always true. In the
result of this removal the disjunction may be converted into a
formula without OR that must be merged into the the AND formula
that contains the disjunction.
The merging of two AND conditions must take into account the
multiple equalities that may be part of each of them.
These multiple equality must be merged and become part of the
and object built as the result of the merge of the AND conditions.
Erroneously the function remove_eq_cond lacked the code that
would merge multiple equalities of the merged AND conditions.
This could lead to confusing situations when at the same AND
level there were two multiple equalities with common members
and the list of equal items contained only some of these
multiple equalities.
This, in its turn, could lead to an incorrect work of the
function substitute_for_best_equal_field when it tried to optimize
ref accesses. This resulted in forming invalid TABLE_REF objects
that were used to build look-up keys when materialized subqueries
were exploited.
This bug in the legacy code could manifest itself in queries with
semi-join materialized subqueries.
When a subquery is materialized all conditions that are imposed
only on the columns belonging to the tables from the subquery
are taken into account.The code responsible for subquery optimizations
that employes subquery materialization makes sure to remove these
conditions from the WHERE conditions of the query obtained after
it has transformed the original query into a query with a semi-join.
If the condition to be removed is an equality condition it could
be added to ON expressions and/or conditions from disjunctive branches
(parts of OR conditions) in an attempt to generate better access keys
to the tables of the query. Such equalities are supposed to be removed
later from all the formulas where they have been added to.
However, erroneously, this was not done in some cases when an ON
expression and/or a disjunctive part of the OR condition could
be converted into one multiple equality. As a result some equality
predicates over columns belonging to the tables of the materialized
subquery remained in the ON condition and/or the a disjunctive
part of the OR condition, and the excuter later, when trying to
evaluate them, returned wrong answers as the values of the fields
from these equalities were not valid.
This happened because any standalone multiple equality (a multiple
equality that are not ANDed with any other predicates) lacked
the information about equality predicates inherited from upper
levels (in particular, inherited from the WHERE condition).
The fix adds a reference to such information to any standalone
multiple equality.
Some queries with the "SELECT ... FROM DUAL" nested subqueries
failed with an assertion on debug builds.
Non-debug builds were not affected.
There were a few different issues with similar assertion
failures on different queries:
1. The first problem was related to the incomplete propagation
of the "non-constant" item status from underlying subquery
items to the outer item tree: in some cases non-constants were
interpreted as constants and evaluated at the preparation stage
(val_int() calls withing fix_fields() etc).
Thus, the default implementation of Item_ref::const_item() from
the Item parent class didn't take into account the "const_item"
status of the referenced item tree -- it used the insufficient
"used_tables() == 0" check instead. This worked in most cases
since our "non-constant" functions like RAND() and SLEEP() set
the RAND_TABLE_BIT in the used table map, so they aren't
non-constant from Item_ref's "point of view". However, the
"SELECT ... FROM DUAL" subquery may have an empty map of used
tables, but at the same time subqueries are never "constant" at
the context analysis stage (preparation, view creation etc).
So, the non-contantness of such subqueries was missed.
Fix: the Item_ref::const_item() function has been overloaded to
take into account both (*ref)->const_item() status and tricky
Item_ref::used_tables() return values, since the only
(*ref)->const_item() call is not enough there.
2. In some cases instead of the const_item() call we check a
value of the Item::with_subselect field to recognize items
with nested subqueries. However, the Item_ref class didn't
propagate this value from the referenced item tree.
Fix: Item::has_subquery() and Item_ref::has_subquery()
functions have been backported from 5.6. All direct
references to the with_subselect fields of nested items have
been with the has_subquery() function call.
3. The Item_func_regex class didn't propagate with_subselect
as well, since it overloads the Item_func::fix_fields()
function with insufficient fix_fields() implementation.
Fix: the Item_func_regex::fix_fields() function has been
modified to gather "constant" statuses from inner items.
4. The Item_func_isnull::update_used_tables() function has
a special branch for the underlying item where the maybe_null
value is false: in this case it marks the Item_func_isnull
as a "const_item" and sets the cached_value to false.
However, the Item_func_isnull::val_int() was not in sync with
update_used_tables(): it didn't take into account neither
const_item_cache nor cached_value for the case of
"args[0]->maybe_null == false optimization".
As far as such an Item_func_isnull has "const_item() == true",
it's ok to call Item_func_isnull::val_int() etc from outer
items on preparation stage. In this case the server tried to
call Item_func_isnull::args[0]->isnull(), and if the args[0]
item contained a nested not-nullable subquery, it failed
with an assertion.
Fix: take the value of Item_func_isnull::const_item_cache into
account in the val_int() function.
5. The auxiliary Item_is_not_null_test class has a similar
optimization in the update_used_tables() function as the
Item_func_isnull class has, and the same issue in the val_int()
function.
In addition to that the Item_is_not_null_test::update_used_tables()
doesn't update the const_item_cache value, so the "maybe_null"
optimization is useless there. Thus, we missed some optimizations
of cases like these (before and after the fix):
< <is_not_null_test>(a),
---
> <cache>(<is_not_null_test>(a)),
or
< having (<is_not_null_test>(a) and <is_not_null_test>(a))
---
> having 1
etc.
Fix: update Item_is_not_null_test::const_item_cache in
update_used_tables() and take in into account in val_int().
from a MERGE view.
The problem was in the lost ability to be null for the table of a left join if it
is a view/derived table.
It hapenned because setup_table_map(), was called earlier then we merged
the view or derived.
Fixed by propagating new maybe_null flag during Item::update_used_tables().
Change in join_outer.test and join_outer_jcl6.test appeared because
IS NULL reported no used tables (i.e. constant) for argument which could not be
NULL and new maybe_null flag was propagated for IS NULL argument (Item_field)
because table the Item_field belonged to changed its maybe_null status.
The problem is a shift operation that is not 64-bit safe.
The consequence is that used tables information for a join with 32 tables
or more will be incorrect.
Fixed by adding a type cast in Item_sum::update_used_tables().
Also used the opportunity to fix some other potential bugs by adding an
explicit type-cast to an integer in a left-shift operation.
Some of them were quite harmless, but was fixed in order to get the same
signed-ness as the other operand of the operation it was used in.
sql/item_cmpfunc.cc
Adjusted signed-ness for some integers in left-shift.
sql/item_subselect.cc
Added type-cast to nesting_map (which is a 32/64 bit type, so
potential bug for deeply nested queries).
sql/item_sum.cc
Added type-cast to nesting_map (32/64-bit type) and table_map
(64-bit type).
sql/opt_range.cc
Added type-cast to ulonglong (which is a 64-bit type).
sql/sql_base.cc
Added type-cast to nesting_map (which is a 32/64-bit type).
sql/sql_select.cc
Added type-cast to nesting_map (32/64-bit type) and key_part_map
(64-bit type).
sql/strfunc.cc
Changed type-cast from longlong to ulonglong, to preserve signed-ness.
Analysis:
The following call stack shows that it is possible to set Item_cache::value_cached, and the relevant value
without setting Item_cache::example.
#0 Item_cache_temporal::store_packed at item.cc:8395
#1 get_datetime_value at item_cmpfunc.cc:915
#2 resolve_const_item at item.cc:7987
#3 propagate_cond_constants at sql_select.cc:12264
#4 propagate_cond_constants at sql_select.cc:12227
#5 optimize_cond at sql_select.cc:13026
#6 JOIN::optimize at sql_select.cc:1016
#7 st_select_lex::optimize_unflattened_subqueries at sql_lex.cc:3161
#8 JOIN::optimize_unflattened_subqueries at opt_subselect.cc:4880
#9 JOIN::optimize at sql_select.cc:1554
The fix is to set Item_cache_temporal::example even when the value is
set directly by Item_cache_temporal::store_packed. This makes the
Item_cache_temporal object consistent.
two tests still fail:
main.innodb_icp and main.range_vs_index_merge_innodb
call records_in_range() with both range ends being open
(which triggers an assert)
We set correct cmp_context during preparation to avoid changing it later by Item_field::equal_fields_propagator.
(see mysql bugs #57135#57692 during merging)
make sure that find_date_time_item() is called before agg_arg_charsets_for_comparison().
optimize Item_func_conv_charset to avoid conversion if no string result is needed
Analysis:
When a subquery that needs a temp table is executed during
the prepare or optimize phase of the outer query, at the end
of the subquery execution all the JOIN_TABs of the subquery
are replaced by a new JOIN_TAB that selects from the temp table.
However that temp table has no corresponding TABLE_LIST.
Once EXPLAIN execution reaches its last phase, it tries to print
the names of the subquery tables through its TABLE_LISTs, but in
the case of this bug there is no such TABLE_LIST (it is NULL),
hence a crash.
Solution:
The fix is to block subquery evaluation inside
Item_func_like::fix_fields and Item_func_like::select_optimize()
using the Item::is_expensive() test.
Analysis:
The problem in the original MySQL bug is that the range optimizer
performs its analysis in a separate MEM_ROOT object that is freed
after the range optimzier is done. During range analysis get_mm_tree
calls Item_func_like::select_optimize, which in turn evaluates its
right argument. In the test case the right argument is a subquery.
In MySQL, subqueries are optimized lazyly, thus the call to val_str
triggers optimization for the subquery. All objects needed by the
subquery plan end up in the temporary MEM_ROOT used by the range
optimizer. When execution ends, the JOIN::cleanup process tries to
cleanup objects of the subquery plan, but all these objects are gone
with the temporary MEM_ROOT. The solution for MySQL is to switch the
mem_root.
In MariaDB with the patch for bug lp:944706, all constant subqueries
that may be used by the optimization process are preoptimized. Therefore
Item_func_like::select_optimize only triggers subquery execution, and
the above problem is not present.
The patch however adds a test whether the evaluated right argument of
the LIKE predicate is expensive. This is consistent with our approach
not to evaluate expensive expressions during optimization.
The patch enables back constant subquery execution during
query optimization after it was disabled during the development
of MWL#89 (cost-based choice of IN-TO-EXISTS vs MATERIALIZATION).
The main idea is that constant subqueries are allowed to be executed
during optimization if their execution is not expensive.
The approach is as follows:
- Constant subqueries are recursively optimized in the beginning of
JOIN::optimize of the outer query. This is done by the new method
JOIN::optimize_constant_subqueries(). This is done so that the cost
of executing these queries can be estimated.
- Optimization of the outer query proceeds normally. During this phase
the optimizer may request execution of non-expensive constant subqueries.
Each place where the optimizer may potentially execute an expensive
expression is guarded with the predicate Item::is_expensive().
- The implementation of Item_subselect::is_expensive has been extended
to use the number of examined rows (estimated by the optimizer) as a
way to determine whether the subquery is expensive or not.
- The new system variable "expensive_subquery_limit" controls how many
examined rows are considered to be not expensive. The default is 100.
In addition, multiple changes were needed to make this solution work
in the light of the changes made by MWL#89. These changes were needed
to fix various crashes and wrong results, and legacy bugs discovered
during development.
Issue/Cause:
Issue is of memory corruption.During optimization phase, pattern to be matched in where
clause, is prepared. This is done in Item_func_concat::val_str() function which forms the
resultant string (tmp_value) and return its pointer. In caller, Item_func_like::fix_fields,
pattern is made to point to this string (tmp_value). In further processing, tmp_value is
getting modified which causes pattern to have changed/wrong values.
Fix:
Allocate its own memroy location in caller, copy value of resultant string (tmp_value)
into that and make pattern to point to that. This makes sure no further changes to
tmp_value will affect pattern.
- Fix equality propagation to work with SJM nests and OR clauses (full descirption of problem and
solution in the comment in the patch)
(The second commit with post-review fixes)
make sure that stored routines are evaluated (that is, de facto - cached) in convert_const_to_int().
revert the fix for lp:806943 because it cannot be repeated anymore.
add few tests for convert_const_to_int()
Fixed Item* Item_equal::get_first(JOIN_TAB *context, Item *field_item) to work correctly in the case where:
- context!= NO_PARTICULAR_TAB, it points to a table within SJ-Materialization nest
- field_item points to an item_equal that has a constant Item_field but does not have any fields
from tables that are within semi-join nests.
Bug#11758543 50756: BIGINT '100' MATCHES 1.001E2
Expressions of the form
BIGINT_COL <compare> <non-integer constant>
should be done either as decimal, or float.
Currently however, such comparisons are done as int,
which means that the constant may be truncated,
and yield false positives/negatives for all queries
where compare is '>' '<' '>=' '<=' '=' '!='.
BIGINT_COL IN <list of contstants>
and
BIGINT_COL BETWEEN <constant> AND <constant>
are also affected.
mysql-test/r/bigint.result:
New tests.
mysql-test/r/func_in.result:
BIGINT <=> string comparison should be done as float,
so a warning for the value 'abc' is appropriate.
mysql-test/t/bigint.test:
New tests.
sql/item_cmpfunc.cc:
In convert_constant_item() we verify that the constant item
can be stored in the given field.
For BIGINT columns (MYSQL_TYPE_LONGLONG) we must verify that the
stored constant value is actually comparable as int,
i.e. that the value was not truncated.
For between: compare as int only if both arguments convert correctly to int.
- equality substitution code was geared towards processing WHERE/ON clauses.
that is, it assumed that it was doing substitions on the code that
= wasn't attached to any particular join_tab yet
= was going to be fed to make_join_select() which would take the condition
apart and attach various parts of it to tables inside/outside semi-joins.
- However, somebody added equality substition for ref access. That is, if
we have a ref access on TBL.key=expr, they would do equality substition in
'expr'. This possibility wasn't accounted for.
- Fixed equality substition code by adding a mode that does equality
substition under assumption that the processed expression will be
attached to a certain particular table TBL.
sql/sql_insert.cc:
CREATE ... IF NOT EXISTS may do nothing, but
it is still not a failure. don't forget to my_ok it.
******
CREATE ... IF NOT EXISTS may do nothing, but
it is still not a failure. don't forget to my_ok it.
sql/sql_table.cc:
small cleanup
******
small cleanup
- Provide fix_after_pullout() function for Item_in_optimizer and other Item_XXX classes (basically, all of them
that have eval_not_null_tables, which means they have special rules for calculating not_null_tables_cache value)
ALL subquery should return TRUE if subquery rowa set is empty independently
of left part. The problem was that Item_func_(eq,ne,gt,ge,lt,le) do not
call execution of second argument if first is NULL no in this case subquery
will not be executed and when Item_func_not_all calls any_value() of the
subquery or aggregation function which report that there was rows. So for
NULL < ALL (SELECT...) result was FALSE instead of TRUE.
Fix is just swapping of arguments of Item_func_(eq,ne,gt,ge,lt,le) (with
changing the operation if it is needed) so that result will be the same
(for examole a < b is equal to b > a). This fix exploit the fact that
first argument will be executed in any case.
- The problem was that the code that made the check whether the subquery is an AND-part of the WHERE
clause didn't work correctly for nested subqueries. In particular, grand-child subquery in HAVING was
treated as if it was in the WHERE, which eventually caused an assert when replace_where_subcondition
looked for the subquery predicate in the WHERE and couldn't find it there.
- The fix: Removed implementation of "thd_marker approach". thd->thd_marker was used to determine the
location of subquery predicate: setup_conds() would set accordingly it when making the
{where|on_expr}->fix_fields(...)
call so that AND-parts of the WHERE/ON clauses can determine they are the AND-parts.
Item_cond_or::fix_fields(), Item_func::fix_fields(), Item_subselect::fix_fields (this one was missed),
and all other items-that-contain-items had to reset thd->thd_marker before calling fix_fields() for
their children items, so that the children can see they are not AND-parts of WHERE/ON.
- The "thd_marker approach" required that a lot of code in different locations maintains correct value of
thd->thd_marker, so it was replaced with:
- The new approach with mark_as_condition_AND_part does not keep context in thd->thd_marker. Instead,
setup_conds() now calls
{where|on_expr}->mark_as_condition_AND_part()
and implementations of that function make sure that:
- parts of AND-expressions get the mark_as_condition_AND_part() call
- Item_in_subselect objects record that they are AND-parts of WHERE/ON
Analysis:
During the first execution of the query through the stored
procedure, the optimization phase calls
substitute_for_best_equal_field(), which calls
Item_in_optimizer::transform(). The latter replaces
Item_in_subselect::left_expr with args[0] via assignment.
In this test case args[0] is an Item_outer_ref which is
created/deallocated for each re-execution. As a result,
during the second execution Item_in_subselect::left_expr
pointed to freed memory, which resulted in a crash.
Solution:
The solution is to use change_item_tree(), so that the
origianal left expression is restored after each execution.
