This commits adds the "materialization" block to the output of
EXPLAIN/ANALYZE FORMAT=JSON when materialized subqueries are involved
into processing. In the case of ANALYZE additional runtime information
is displayed, such as:
- chosen strategy of materialization
- number of partial match/index lookup loops
- sizes of partial match buffers
from HAVING
The bug is caused by refixing of the constant subquery in pushdown from
HAVING into WHERE optimization.
Similarly to MDEV-29363 in the problematic query two references of the
constant subquery are used. After the pushdown one of the references of the
subquery is pushed into WHERE-clause and the second one remains as the part
of the HAVING-clause.
Before the represented fix, the constant subquery reference that was going to
be pushed into WHERE was cleaned up and fixed. That caused the changes of
the subquery itself and, therefore, changes for the second reference that
remained in HAVING. These changes caused a crash.
To fix this problem all constant objects that are going to be pushed into
WHERE should be marked with an IMMUTABLE_FL flag. Objects marked with this
flag are not cleaned up or fixed in the pushdown optimization.
Approved by Igor Babaev <igor@mariadb.com>
The crash is caused by the attempt to refix the constant subquery during
pushdown from HAVING into WHERE optimization.
Every condition that is going to be pushed into WHERE clause is first
cleaned up, then refixed. Constant subqueries are not cleaned or refixed
because they will remain the same after refixing, so this complicated
procedure can be omitted for them (introduced in MDEV-21184).
Constant subqueries are marked with flag IMMUTABLE_FL, that helps to miss
the cleanup stage for them. Also they are marked as fixed, so refixing is
also not done for them.
Because of the multiple equality propagation several references to the same
constant subquery can exist in the condition that is going to be pushed
into WHERE. Before this patch, the problem appeared in the following way.
After the first reference to the constant subquery is processed, the flag
IMMUTABLE_FL for the constant subquery is disabled.
So, when the second reference to this constant subquery is processed, the
flag is already disabled and the subquery goes through the procedure of
cleaning and refixing. That causes a crash.
To solve this problem, IMMUTABLE_FL should be disabled only after all
references to the constant subquery are processed, so after the whole
condition that is going to be pushed is cleaned up and refixed.
Approved by Igor Babaev <igor@maridb.com>
Having Item_func_not items in item trees breaks assumptions during the
optimization phase about transformation possibilities in fix_fields().
Remove Item_func_not by extending normalization during parsing.
Reviewed by Oleksandr Byelkin (sanja@mariadb.com)
The problem is that the first execution of the prepared statement makes
a permanent optimization of converting the LEFT JOIN to an INNER JOIN.
This is based on the assumption that all the user parameters (?) are
always constants and that parameters to Item_cond() will not change value
from true and false between different executions.
(The example was using IS NULL, which will change value if parameter
depending on if the parameter is NULL or not).
The fix is to change Item_cond::fix_fields() and
Item_cond::eval_not_null_tables() to not threat user parameters as
constants. This will ensure that we don't do the LEFT_JOIN -> INNER
JOIN conversion that causes problems.
There is also some things that needs to be improved regarding
calculations of not_null_tables_cache as we get a different value for
WHERE 1 or t1.a=1
compared to
WHERE t1.a= or 1
Changes done:
- Mark Item_param with the PARAM flag to be able to quickly check
in Item_cond::eval_not_null_tables() if an item contains a
prepared statement parameter (just like we check for stored procedure
parameters).
- Fixed that Item_cond::not_null_tables_cache is not depending on
order of arguments.
- Don't call item->eval_const_cond() for items that are NOT on the top
level of the WHERE clause. This removed a lot of unnecessary
warnings in the test suite!
- Do not reset not_null_tables_cache for not top level items.
- Simplified Item_cond::fix_fields by calling eval_not_null_tables()
instead of having duplication of all the code in
eval_not_null_tables().
- Return an error if Item_cond::fix_field() generates an error
The old code did generate an error in some cases, but not in all
cases.
- Fixed all handling of the above error in make_cond_for_tables().
The error handling by the callers did not exists before which
could lead to asserts in many different places in the old code).
- All changes in sql_select.cc are just checking the return value of
fix_fields() and make_cond_for_tables() and returning an error
value if fix_fields() returns true or make_cond_for_tables()
returns NULL and is_error() is set.
- Mark Item_cond as const_item if all arguments returns true for
can_eval_in_optimize().
Reviewer: Sergei Petrunia <sergey@mariadb.com>
The main difference in code path between EQ_REF and REF is that for
REF we have to do an extra read_next on the index to check that there
is no more matching rows.
Before this patch we added a preference of EQ_REF by ensuring that REF
would always estimate to find at least 2 rows.
This patch adds the cost of the extra key read_next to REF access and
removes the code that limited REF to at least 2 rows. For some queries
this can have a big effect as the total estimated rows will be halved
for each REF table with 1 rows.
multi_range cost calculations are also changed to take into account
the difference between EQ_REF and REF.
The effect of the patch to the test suite:
- About 80 test case changed
- Almost all changes where for EXPLAIN where estimated rows for REF
where changed from 2 to 1.
- A few test cases using explain extended had a change of 'filtered'.
This is because of the estimated rows are now closer to the
calculated selectivity.
- A very few test had a change of table order.
This is because the change of estimated rows from 2 to 1 or the small
cost change for REF
(main.subselect_sj_jcl6, main.group_by, main.dervied_cond_pushdown,
main.distinct, main.join_nested, main.order_by, main.join_cache)
- No key statistics and the estimated rows are now smaller which cased
estimated filtering to be lower.
(main.subselect_sj_mat)
- The number of total rows are halved.
(main.derived_cond_pushdown)
- Plans with 1 row changed to use RANGE instead of REF.
(main.group_min_max)
- ALL changed to REF
(main.key_diff)
- Key changed from ref + index_only to PRIMARY key for InnoDB, as
OPTIMIZER_ROW_LOOKUP_COST + OPTIMIZER_ROW_NEXT_FIND_COST is smaller than
OPTIMIZER_KEY_LOOKUP_COST + OPTIMIZER_KEY_NEXT_FIND_COST.
(main.join_outer_innodb)
- Cost changes printouts
(main.opt_trace*)
- Result order change
(innodb_gis.rtree)
MDEV-28082 Crash when using HAVING with IS NULL predicate in an equality
These bugs have been fixed by the patch for MDEV-26402.
Only test cases are added.
When doing condition pushdown from HAVING into WHERE,
Item_equal::create_pushable_equalities() calls
item->set_extraction_flag(IMMUTABLE_FL) for constant items.
Then, Item::cleanup_excluding_immutables_processor() checks for this flag
to see if it should call item->cleanup() or leave the item as-is.
The failure happens when a constant item has a non-constant one inside it,
like:
(tbl.col=0 AND impossible_cond)
item->walk(cleanup_excluding_immutables_processor) works in a bottom-up
way so it
1. will call Item_func_eq(tbl.col=0)->cleanup()
2. will not call Item_cond_and->cleanup (as the AND is constant)
This creates an item tree where a fixed Item has an un-fixed Item inside
it which eventually causes an assertion failure.
Fixed by introducing this rule: instead of just calling
item->set_extraction_flag(IMMUTABLE_FL);
we call Item::walk() to set the flag for all sub-items of the item.
MDEV-27036: repeated "table" key resolve for print_explain_json
MDEV-27036: duplicated keys in best_access_path
MDEV-27036: Explain_aggr_filesort::print_json_members: resolve duplicated "filesort" member in Json object
MDEV-27036: Explain_basic_join::
print_explain_json_interns fixed start_dups_weedout case for main.explain_json test
with condition_pushdown_from_having
This bug could manifest itself for queries with GROUP BY and HAVING clauses
when the HAVING clause was a conjunctive condition that depended
exclusively on grouping fields and at least one conjunct contained an
equality of the form fld=sq where fld is a grouping field and sq is a
constant subquery.
In this case the optimizer tries to perform a pushdown of the HAVING
condition into WHERE. To construct the pushable condition the optimizer
first transforms all multiple equalities in HAVING into simple equalities.
This has to be done for a proper processing of the pushed conditions
in WHERE. The multiple equalities at all AND/OR levels must be converted
to simple equalities because any multiple equality may refer to a multiple
equality at the upper level.
Before this patch the conversion was performed like this:
multiple_equality(x,f1,...,fn) => x=f1 and ... and x=fn.
When an equality item for x=fi was constructed both the items for x and fi
were cloned. If x happened to be a constant subquery that could not be
cloned the conversion failed. If the conversions of multiple equalities
previously performed had succeeded then the whole condition became in an
inconsistent state that could cause different failures.
The solution provided by the patch is:
1. to use a different conversion rule if x is a constant
multiple_equality(x,f1,...,fn) => f1=x and f2=f1 and ... and fn=f1
2. not to clone x if it's a constant.
Such conversions cannot fail and besides the result of the conversion
preserves the equivalence of f1,...,fn that can be used for other
optimizations.
This patch also made sure that expensive predicates are not pushed from
HAVING to WHERE.
The bug appears because not all conditions are found to be knowingly
true or false in WHERE after HAVING pushdown optimization.
Impossible WHERE can be found much earlier compared with how it is done now.
To fix it and_new_conditions_to_optimized_cond() is changed.
When pushing a condition from HAVING into WHERE the function
st_select_lex::pushdown_from_having_into_where() transforms column
references in the pushed condition then performs cleanup of
items of the condition and finally calls fix_fields() for the condition
items. The cleanup is performed by a call of the method walk() with
cleanup_processor as the first parameter. Unfortunately this sequence
of calls does not work if the condition contains cached items, because
fix_fields() cannot go through Item_cache items and this leaves
underlying items unfixed.
The solution of this problem used in this patch is just does not allow
to process Item_cache objects when performing cleanup of the pushed
condition. In order to let the traversal procedure walk() not to process
Item_cache objects the third parameter of the used call of walk()
is set to a fictitious pointer (void *) 1. And Item_cache::walk() is
changed to prevent any action when it gets such value as the third
parameter.
Temporary table is defined with the view field in HAVING.
Item_direct_view_ref for this field is dropped and that causes error.
To fix it Item_direct_view_ref::remove_item_direct_ref() is added.
The bug occurs because of the wrong pushdown of constant function
defined with subquery from HAVING into WHERE. Subqueries can't be
pushed into WHERE.
To fix it with_subquery() call is added to check if the function contains
subquery.
This bug is caused by pushdown from HAVING into WHERE.
It appears because condition that is pushed wasn't fixed.
It is also discovered that condition pushdown from HAVING into
WHERE is done wrong. There is no need to build clones for some
conditions that can be pushed. They can be simply moved from HAVING
into WHERE without cloning.
build_pushable_cond_for_having_pushdown(),
remove_pushed_top_conjuncts_for_having() methods are changed.
It is found that there is no transformation made for fields of
pushed condition.
field_transformer_for_having_pushdown transformer is added.
New tests are added. Some comments are changed.
Optimized the code that removed multiple equalities pushed from HAVING
into WHERE. Now this removal is postponed until all multiple equalities
are eliminated in substitute_for_best_equal_field().
Condition can be pushed from the HAVING clause into the WHERE clause
if it depends only on the fields that are used in the GROUP BY list
or depends on the fields that are equal to grouping fields.
Aggregate functions can't be pushed down.
How the pushdown is performed on the example:
SELECT t1.a,MAX(t1.b)
FROM t1
GROUP BY t1.a
HAVING (t1.a>2) AND (MAX(c)>12);
=>
SELECT t1.a,MAX(t1.b)
FROM t1
WHERE (t1.a>2)
GROUP BY t1.a
HAVING (MAX(c)>12);
The implementation scheme:
1. Extract the most restrictive condition cond from the HAVING clause of
the select that depends only on the fields that are used in the GROUP BY
list of the select (directly or indirectly through equalities)
2. Save cond as a condition that can be pushed into the WHERE clause
of the select
3. Remove cond from the HAVING clause if it is possible
The optimization is implemented in the function
st_select_lex::pushdown_from_having_into_where().
New test file having_cond_pushdown.test is created.
