To correctly decide which predicates can be evaluated with a given table
the optimizer must know the exact set of tables that a predicate depends
on. If that mask is too wide (refer to non-existing tables) the optimizer
can erroneously skip a predicate.
One such case of wrong table usage mask were the aggregate functions.
The have a all-1 mask (meaning depend on all tables, including non-existent
ones).
Fixed by making a real used_tables mask for the aggregates. The mask is
constructed in the following way :
1. OR the table dependency masks of all the arguments of the aggregate.
2. If all the arguments of the function are from the local name resolution
context and it is evaluated in the same name resolution
context where it is referenced all the tables from that name resolution
context are OR-ed to the dependency mask. This is to denote that an
aggregate function depends on the number of rows it processes.
3. Handle correctly the case of an aggregate function optimization (such that
the aggregate function can be pre-calculated and made a constant).
Made sure that an aggregate function is never a constant (unless subject of a
specific optimization and pre-calculation).
One other flaw was revealed and fixed in the process : references were
not calling the recalculation method for used_tables of their targets.
mysql-test/r/subselect3.result:
Bug #24484: test case
mysql-test/t/subselect3.test:
Bug #24484: test case
sql/item.h:
Bug #24484: Item_ref must update the used tables.
sql/item_sum.cc:
Bug #24484: correct calculation of used_tables for aggregates.
sql/item_sum.h:
Bug #24484: correct calculation of used_tables for aggregates.
sql/opt_range.cc:
Bug #24484: fixed ref resolution in loose index scan
sql/sql_base.cc:
Bug #24484: moved counting of leaf tables inside
setup_tables_and_check_access.
sql/sql_class.h:
Bug #24484: changed table count to more narrow type.
sql/sql_insert.cc:
Bug #24484: moved counting of leaf tables inside
setup_tables_and_check_access. Substract the first
table (and its subtables) of an INSERT statement
from leaf_count.
sql/sql_select.cc:
Bug #24484: correct check for aggregates
Objects of the classes Item_func_is_not_null_test and Item_func_trig_cond
must be transparent for the method Item::split_sum_func2 as these classes
are pure helpers. It means that the method Item::split_sum_func2 should
look at those objects as at pure wrappers.
mysql-test/r/subselect3.result:
Added a test case for bug #24420.
mysql-test/t/subselect3.test:
Added a test case for bug #24420.
- Make the code produce correct result: use an array of triggers to turn on/off equalities for each
compared column. Also turn on/off optimizations based on those equalities.
- Make EXPLAIN output show "Full scan on NULL key" for tables for which we switch between
ref/unique_subquery/index_subquery and ALL access.
- index_subquery engine now has HAVING clause when it is needed, and it is
displayed in EXPLAIN EXTENDED
- Fix incorrect presense of "Using index" for index/unique-based subqueries (BUG#22930)
// bk trigger note: this commit refers to BUG#24127
mysql-test/r/ndb_subquery.result:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Updated test results (checked)
mysql-test/r/subselect.result:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Updated test results (checked)
mysql-test/r/subselect2.result:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Updated test results (checked)
mysql-test/r/subselect3.result:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Testcases
mysql-test/t/subselect3.test:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Testcases
sql/item_cmpfunc.cc:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- For row-based IN subqueries, use one flag per each column. Set the flags appropriately before
running the subquery.
sql/item_cmpfunc.h:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Added Item_func_trig_cond::get_triv_var()
sql/item_subselect.cc:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Item_subselect::exec() and subselect_*_engine::exec() don't have parameter
anymore - now Item_subselect owns the pushed down predicates guard flags.
- A correct set of conditional predicates is now pushed into row-based IN
subquery.
- select_indexsubquery_engine now has "HAVING clause" (needed for correct query
results), and it is shown in EXPLAIN EXTENDED
sql/item_subselect.h:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Item_subselect::exec() and subselect_*_engine::exec() don't have parameter
anymore - now Item_subselect owns the pushed down predicates guard flags.
- A correct set of conditional predicates is now pushed into row-based IN
subquery.
- select_indexsubquery_engine now has "HAVING clause" (needed for correct query
results), and it is shown in EXPLAIN EXTENDED
sql/mysql_priv.h:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Added "in_having_cond" special Item name
sql/mysqld.cc:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Added "in_having_cond" special Item name
sql/sql_lex.h:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
sql/sql_select.cc:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Make "ref" analyzer be able to work with conditional equalities
- Fix subquery optimization code to match the changes in what kinds of
conditions are pushed down into subqueries
- Fix wrong EXPLAIN output in some queries with subquery (BUG#22390)
sql/sql_select.h:
BUG#24127: wrong result for (null,not-null) IN (SELECT a,b ...)
- Make "ref" analyzer be able to work with conditional equalities
- Fix wrong EXPLAIN output in some queries with subquery (BUG#22390)
When transforming "oe IN (SELECT ie ...)" wrap the pushed-down predicates
iff "oe can be null", not "ie can be null".
The fix doesn't cover row-based subqueries, those will be fixed in #24127.
mysql-test/r/subselect.result:
BUG#24085: Wrong result for NULL IN (SELECT not_null_val FROM ...)
Update the test results (checked)
mysql-test/r/subselect3.result:
BUG#24085: Wrong result for NULL IN (SELECT not_null_val FROM ...)
- Testcase
mysql-test/t/subselect3.test:
BUG#24085: Wrong result for NULL IN (SELECT not_null_val FROM ...)
- Testcase
sql/item_subselect.cc:
BUG#24085: Wrong result for NULL IN (SELECT not_null_val FROM ...)
When transforming "oe IN (SELECT ie ...)" we should make special
provisions (wrap the pushed predicates) if we can encounter
NULL IN (SELECT ...), i.e. when oe->maybe_null. The code was checking
for ie->maybe_null instead, fixed it for single value based subqueries.
Row-based subqueries (e.g. (a,b) IN (SELECT c,d ...)) are not fixed
because they won't produce correct results for several other reasons
(filed as #24085)
Evaluate "NULL IN (SELECT ...)" in a special way: Disable pushed-down
conditions and their "consequences":
= Do full table scans instead of unique_[index_subquery] lookups.
= Change appropriate "ref_or_null" accesses to full table scans in
subquery's joins.
Also cache value of NULL IN (SELECT ...) if the SELECT is not correlated
wrt any upper select.
mysql-test/r/subselect.result:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Updated test results
sql/item.h:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Added comments
sql/item_cmpfunc.cc:
BUG#8804: wrong results for NULL IN (SELECT ...):
Made Item_in_optimizer to:
- cache the value of "NULL IN (uncorrelated select)"
- Turn off pushed-down predicates when evaluating "NULL IN (SELECT ...)"
sql/item_cmpfunc.h:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Made Item_in_optimizer cache the value of "NULL IN (uncorrelated select)"
- Added comments
sql/item_subselect.cc:
BUG#8804: wrong results for NULL IN (SELECT ...):
- When needed, wrap the predicates we push into subquery into an
Item_func_trig_cond so we're able to turn them off when evaluating
NULL IN (SELECT ...).
- Added code to evaluate NULL IN (SELECT ...) in a special way:
= In [unique_]index_subquery, do full table scan to see if there
are any rows.
= For other subqueries, change ref[_or_null] to ALL if the
ref[_or_null] was created from pushed-down predicate.
sql/item_subselect.h:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Added Item_subselect::is_correlated
- Added comments
sql/records.cc:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Make rr_sequential() non-static
sql/sql_lex.cc:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Added st_select_lex::is_correlated and Item_subselect::is_correlated.
sql/sql_lex.h:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Added st_select_lex::is_correlated
sql/sql_select.cc:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Added KEY_FIELD::outer_ref to keep track of which ref accesses are
created from predicates that were pushed down into the subquery.
sql/sql_select.h:
BUG#8804: wrong results for NULL IN (SELECT ...):
- Added KEYUSE::outer_ref
mysql-test/r/subselect3.result:
New BitKeeper file ``mysql-test/r/subselect3.result''
mysql-test/t/subselect3.test:
New BitKeeper file ``mysql-test/t/subselect3.test''
