This makes the keys smaller (no row pointer) and gives us proper errors if we
use the table wrongly.
sql/sql_select.cc:
Use NO_RECORD for tables that doesn't need row data.
storage/maria/Makefile.am:
Added ma_norec.c
storage/maria/ma_check.c:
Added support for NO_RECORD record format (don't store any row data)
storage/maria/ma_norec.c:
Added support for NO_RECORD record format
storage/maria/ma_open.c:
Added support for NO_RECORD record format
storage/maria/ma_search.c:
Added support for 0 size row pointers (used with NO_RECORD)
storage/maria/ma_test1.c:
Added testing of NO_RECORD record format.
storage/maria/maria_chk.c:
Added support for NO_RECORD
storage/maria/maria_def.h:
Added support for NO_RECORD
storage/maria/unittest/ma_test_all-t:
Added testing of NO_RECORD record format
Made sure that the optimal fields are used by TABLE_REF objects
when building index access keys to joined tables.
Fixed a bug in the template function that sorts the elements of
a list using the bubble sort algorithm. The bug caused poor
performance of the function. Also added an optimization that
skips comparison with the most heavy elements that has been
already properly placed in the list.
Made the comparison of the fields belonging to the same Item_equal
more granular: fields belonging to the same table are also ordered
according to some rules.
Date: Mon, 01 Nov 2010 15:15:25 -0000
3272 Roy Lyseng 2010-11-01
Bug#52068: Optimizer generates invalid semijoin materialization plan
When MaterializeScan semijoin strategy was used and there were one
or more outer dependent tables before the semijoin tables, the scan
over the materialized table was not properly reset for each row of
the prefix outer tables.
Example: suppose we have a join order:
ot1 SJ-Mat-Scan(it2 it3) ot4
Notice that this is called a MaterializeScan, even though there is an
outer table ahead of the materialized tables. Usually a MaterializeScan
has the outer tables after the materialized table, but this is
a special (but legal) case with outer dependent tables both before and
after the materialized table.
For each qualifying row from ot1, a new scan over the materialized
table must be set up. The code failed to do that, so all scans after
the first one returned zero rows from the materialized table.
Analysis:
The assert failed because st_select_lex::print() was called for subqueries
as follows:
Item_subselect::print() ->
subselect_single_select_engine::print() -> st_select_lex::print()
It was Item_subselect::fix_fields() that set the thd by calling set_thd(),
so when this print() was called before fix_fields(), subselect_engine::thd
was NULL.
Solution:
The patch makes all constructors of all subselect_engine classes to take
a THD parameter. The default subselect_single_select_engine engine is created
early during parse time, in the Item_subselect::init call, so we pass the
correct THD object already at this point.
- Fixed problem with oqgraph and 'make dist'
Note that after this merge we have a problem show in join_outer where we examine too many rows in one specific case (related to BUG#57024).
This will be fixed when mwl#128 is merged into 5.3.
Changed rows_read and rows_sent status variables to be longlong (to avoid compiler warnings)
sql/item_func.cc:
Fixed wrong usage of alias
sql/item_subselect.cc:
Changed buffer size to ulonglong to be able detect buffers bigger than size_t
sql/item_subselect.h:
Changed buffer size to ulonglong to be able detect buffers bigger than size_t
sql/multi_range_read.cc:
Fixed compiler warning by using correct type for function argument
sql/mysqld.cc:
Changed rows_read and rows_sent status variables to be longlong
sql/opt_subselect.h:
Fixed compiler warning by using correct type for function argument
sql/sql_class.cc:
Changed rows_read and rows_sent status variables to be longlong
sql/sql_class.h:
Changed rows_read and rows_sent status variables to be longlong
Changed max_nulls_in_row to uint as this is number of columns in row.
This fixed some compiler warnings.
sql/sql_select.cc:
Added casts to avoid compiler warnings
storage/heap/ha_heap.cc:
Initilize different types separate
storage/oqgraph/ha_oqgraph.cc:
Fixed argument to store(longlong) to avoid compiler warnings
- Changed Cached_item_field not copy data for fields with NULL value
- In key_copy() and key_restore() don't copy data for fields with NULL value
Fixed code to avoid valgrind warnings
- Use c_ptr_safe instead of c_ptr()
Removed "QQ" from comments (QQ was ment to be used for internal comments that should be removed before pushing)
Fixed wrong alias used (from previous patch)
sql/event_db_repository.cc:
Update testing if event table is valid (to avoid valgrind errors)
sql/ha_partition.cc:
m_ordered_scan_ongoing was not initialized
Reset null bits in record to avoid valgrind errors
sql/handler.h:
Added flag if storage engine will write row verbatim and the row contains varchar or null fields
(in which case we must clear the row to avoid valgrind warnings)
sql/item_buff.cc:
Changed Cached_item_field not copy data for fields with NULL value
(Optimization and avoids valgrind warnings)
sql/item_func.cc:
c_ptr() -> c_ptr_safe()
sql/key.cc:
In key_copy() and key_restore() don't copy data for fields with NULL value
sql/opt_range.cc:
c_ptr() -> c_ptr_safe()
sql/sql_base.cc:
Added TRASH() to table->record[0] to find out if we access not initialzed data.
Initialize null_bytes to:
- Get consistent tests
- Ensure we don't get valgrind warnings for null fields (as we may only update a couple of bits in a byte)
sql/sql_class.cc:
Removed "QQ" from comments
sql/sql_insert.cc:
Initialize row to default values if we are using valgrind and row will be copied verbatim to disk in storage engine.
sql/sql_load.cc:
QQ -> TODO
sql/sql_parse.cc:
Removed old not used code marked QQ and withing "#ifdef REMOVED"
sql/sql_select.cc:
QQ -> TODO
Initialize some variables that was used uninitialized
Added DBUG_ASSERT() to find out if thd was not properly initialized for sub queries
sql/sql_test.cc:
Fixed format for printing to DBUG file
Fixed wrong alias used (from previous patch)
sql/sql_trigger.h:
QQ -> TODO
sql/table.cc:
QQ -> TODO
storage/maria/ha_maria.cc:
Mark table with HA_RECORD_MUST_BE_CLEAN_ON_WRITE, if row is written verbatim to disk and contains varchar or null fields.
storage/maria/ma_open.c:
Added flags if table has varchar or null fields
storage/maria/maria_def.h:
Added flags if table has varchar or null fields
storage/myisam/ha_myisam.cc:
Mark table with HA_RECORD_MUST_BE_CLEAN_ON_WRITE, if row is written verbatim to disk and contains varchar or null fields.
storage/myisam/mi_open.c:
Fixed memory overrun bug when using fulltext keys
storage/xtradb/row/row0sel.c:
Removed initialization of null bits. (not needed anymore)
Fix MySQL BUG#52344 - Subquery materialization: Assertion if subquery in on-clause of outer join
Original fix and comments from Oysten, adjusted for the different
subquery optimization in MariaDB.
"
Problem: If tables of an outer join are constant tables,
the associated on-clause will be evaluated in the optimization
phase. If the on-clause contains a query that is to be
executed with subquery materialization, this will not work
since the infrastructure for such execution is not yet set up.
Solution: Do not evaluate on-clause in optimization phase if
is_expensive() returns true for this clause. This is how the
problem is currently avoided for where-clauses. This works
because, Item_in_subselect::is_expensive_processor returns true
if query is to be executed with subquery materialization.
"
In addition, after MWL#89, in MariaDB if the IN-EXISTS strategy
is chosen, the in-to-exists predicates are insterted after
join_read_const_table() is called, resulting in evaluation of
the subquery without the in-to-exists predicates.
Open issues:
- A better fix for #57688; Igor is working on this
- Test failure in index_merge_innodb.test ; Igor promised to look at this
- Some Innodb tests fails (need to merge with latest xtradb) ; Kristian promised to look at this.
- Failing tests: innodb_plugin.innodb_bug56143 innodb_plugin.innodb_bug56632 innodb_plugin.innodb_bug56680 innodb_plugin.innodb_bug57255
- Werror is disabled; Should be enabled after merge with xtradb.
- Changed TABLE->alias to String to get fewer reallocs when alias are used.
- Preallocate some buffers
Changed some String->c_ptr() -> String->ptr() when \0 is not needed.
Fixed wrong usage of String->ptr() when we need a \0 terminated string.
Use my_strtod() instead of my_atof() to avoid having to add \0 to string.
c_ptr() -> c_ptr_safe() to avoid warnings from valgrind.
zr
sql/event_db_repository.cc:
Update usage of TABLE->alias
sql/event_scheduler.cc:
c_ptr() -> c_ptr_safe()
sql/events.cc:
c_ptr() -> ptr() as \0 was not needed
sql/field.cc:
Update usage of TABLE->alias
sql/field.h:
Update usage of TABLE->alias
sql/ha_partition.cc:
Update usage of TABLE->alias
sql/handler.cc:
Update usage of TABLE->alias
Fixed wrong usage of str.ptr()
sql/item.cc:
Fixed error where code wrongly assumed string was \0 terminated.
sql/item_func.cc:
c_ptr() -> c_ptr_safe()
Update usage of TABLE->alias
sql/item_sum.h:
Use my_strtod() instead of my_atof() to avoid having to add \0 to string
sql/lock.cc:
Update usage of TABLE->alias
sql/log.cc:
c_ptr() -> ptr() as \0 was not needed
sql/log_event.cc:
c_ptr_quick() -> ptr() as \0 was not needed
sql/opt_range.cc:
ptr() -> c_ptr() as \0 is needed
sql/opt_subselect.cc:
Update usage of TABLE->alias
sql/opt_table_elimination.cc:
Update usage of TABLE->alias
sql/set_var.cc:
ptr() -> c_ptr() as \0 is needed
c_ptr() -> c_ptr_safe()
sql/sp.cc:
c_ptr() -> ptr() as \0 was not needed
sql/sp_rcontext.cc:
Update usage of TABLE->alias
sql/sql_base.cc:
Preallocate buffers
Update usage of TABLE->alias
sql/sql_class.cc:
Fix arguments to sprintf() to work even if string is not \0 terminated
sql/sql_insert.cc:
Update usage of TABLE->alias
c_ptr() -> ptr() as \0 was not needed
sql/sql_load.cc:
Preallocate buffers
Trivial optimizations
sql/sql_parse.cc:
Trivial optimization
sql/sql_plugin.cc:
c_ptr() -> ptr() as \0 was not needed
sql/sql_select.cc:
Update usage of TABLE->alias
sql/sql_show.cc:
Update usage of TABLE->alias
sql/sql_string.h:
Added move() function to move allocated memory from one object to another.
sql/sql_table.cc:
Update usage of TABLE->alias
c_ptr() -> c_ptr_safe()
sql/sql_test.cc:
ptr() -> c_ptr_safe()
sql/sql_trigger.cc:
Update usage of TABLE->alias
c_ptr() -> c_ptr_safe()
sql/sql_update.cc:
Update usage of TABLE->alias
sql/sql_view.cc:
ptr() -> c_ptr_safe()
sql/sql_yacc.yy:
ptr() -> c_ptr()
sql/table.cc:
Update usage of TABLE->alias
sql/table.h:
Changed TABLE->alias to String to get fewer reallocs when alias are used.
storage/federatedx/ha_federatedx.cc:
Use c_ptr_safe() to ensure strings are \0 terminated.
storage/maria/ha_maria.cc:
Update usage of TABLE->alias
storage/myisam/ha_myisam.cc:
Update usage of TABLE->alias
storage/xtradb/row/row0sel.c:
Ensure that null bits in record are properly reset.
(Old code didn't work as row_search_for_mysql() can be called twice while reading fields from one row.
The condition that was supposed to check whether a join table
is an inner table of a nested outer join or semi-join was not
quite correct in the code of the function check_join_cache_usage.
That's why some queries with nested outer joins triggered
an assertion failure.
Encapsulated this condition in the new method called
JOIN_TAB::is_nested_inner and provided a proper code for it.
Also corrected a bug in the code of check_join_cache_usage()
that caused a downgrade of not first join buffers of the
level 5 and 7 to level 4 and 6 correspondingly.
When pushing the condition for a table in the function
JOIN_TAB::make_scan_filter the optimizer must not push
conditions from WHERE if the table is some inner table
of an outer join..
The condition over outer tables extracted from the on expression
for a outer join must be ANDed to the condition pushed to the
first inner table of this outer join only.
Nested outer joins cannot use flat join buffers. So if join_cache_level
is set to 1 then any join algorithm employing join buffers cannot be used
for nested outer joins.
A non-incremental join buffer cannot be used for inner tables of nested
outer joins. That's why when join_cache_level is set to 7 it must
be downgraded to level 6 for the inner tables of nested outer joins.
For the same reason with join_cache_level set to 3 no join buffer is
used for the inner tables of outer joins (we could downgrade it to
level 2, but this level does not support ref access).
Miscalculation of the minimum possible buffer size could trigger
an assert in JOIN_CACHE_HASHED::put_record when if join_buffer_size
was set to the values that is less than the length of one record to
stored in the join buffer.
It happened due to the following mistakes:
- underestimation of space needed for a key in the hash table
(we have to take into account that hash table can have more
buckets than the expected number of records).
- the value of maximum total length of all records stored in
the join buffer was not saved in the field max_used_fieldlength
by the function calc_used_field_length.
The pushdown condition for the sorted table in a query can be complemented
by the conditions from HAVING. This transformation is done in JOIN::exec
pretty late after the original pushdown condition have been saved in the
field pre_idx_push_select_cond for the sorted table. So this field must
be updated after the inclusion of the condition from HAVING.
Avoid doing reallocs
Prealloc some strings / provide extension allocation size to some strings
This gave a 25 % speedup in some mysql-test-run tests.
mysys/safemalloc.c:
More DBUG_PRINT
sql/net_serv.cc:
Make all mallocs() look the similar. (just-for-safety fix)
sql/protocol.cc:
Ensure that communication packet buffer is allocated.
(It's freed by stored precedures and some DLL statements)
sql/sp.cc:
Fixed valgrind warning
sql/sql_select.cc:
Set extent allocation for buffer that has a lot of append() calls.
sql/sql_show.cc:
Fixed wrong usage of string buffer. Old code worked in test suite 'just-by-chance'
sql/sql_string.cc:
Call realloc_with_extra_if_needed() in append() functions.
sql/sql_string.h:
Added 'extra_alloc' member, to specify chunck size for realloc().
extra_alloc is addaptive to catch cases where preallocation of buffers is not done properly.
Simplified free() to allow compiler to optimize things better (and to keep things consistent).
Fixed shrink() to take into account the extra memory added to the Alloced_length in realloc(). This saves us a realloc() per query.
sql/sql_test.cc:
Set extent allocation for buffer that has a lot of append() calls.
sql/table.cc:
Set extent allocation for buffer that has a lot of append() calls.
try to find a match in the join buffer. It makes sense to check for a match
only those records satisfying WHERE/ON conditions that can be pushed to
the scanned table. It allows us to discard early some join candidates.
Such pushdown conditions were built when BNL join algorithm was employed,
but for they were not built when BNLH algorithm was used.
The patch removes this shortcoming.
plans or wrong results due to the fact that JOIN_CACHE functions
ignored the possibility of interleaving materialized semijoin
tables with tables whose records were stored in join buffers.
This fixes would become mostly unnecessary if the new code of
mwl 90 was merged into 5.3 right now.
Yet the fix the code of optimize_wo_join_buffering was needed
in any case.
- Make optimize_wo_join_buffering() handle cases where position->records_read=0 (this
happens for outer joins that have constant tables inside them). The number of
0 is not correct (should be 1 because outer join will produce at least a NULL-complemented
record) but for now we just make it work with incorrect number.
Prohibited to use hash join algorithm BNLH if join attributes
need non-binary collations. It has to be done because BNLH does
not support join for such attributes yet.
Later this limitations will be lifted.
Changed default collations for the schemes of some test cases
to preserve the old execution plans.
Lines below which were added in the patch for Bug#56814 cause this crash:
+ if (table->table)
+ table->table->maybe_null= FALSE;
Consider following test case:
--
CREATE TABLE t1(f1 INT NOT NULL);
INSERT INTO t1 VALUES (16777214),(0);
SELECT COUNT(*) FROM t1 LEFT JOIN t1 t2
ON 1 WHERE t2.f1 > 1 GROUP BY t2.f1;
DROP TABLE t1;
--
We set TABLE::maybe_null to FALSE for t2 table
and in create_tmp_field() we create appropriate tmp table field
using create_tmp_field_from_item() function instead of
create_tmp_field_from_field. As a result we have
LONGLONG field. As we have GROUP BY clause we calculate
group buffer length, see calc_group_buffer().
Item from group list which is used for calculation
refer to the field from real tables and have LONG type.
So group buffer length become insufficient for storing of
LONGLONG value. It leads to overwriting of wrong memory
area in do_field_int() function which is called from
end_update().
After some investigation I found out that
create_tmp_field_from_item() is used only for OLAP
grouping and can not be used for common grouping
as it could be an incompatibility between tmp
table fields and group buffer length.
We can not remove create_tmp_field_from_item() call from
create_tmp_field as OLAP needs it and we can not use this
function for common grouping. So we should remove setting
TABLE::maybe_null to FALSE from simplify_joins().
In this case we'll get wrong behaviour of
list_contains_unique_index() back. To fix it we
could use Field::real_maybe_null() check instead of
Field::maybe_null() and add addition check of
TABLE_LIST::outer_join.
mysql-test/r/group_by.result:
test case
mysql-test/r/join_outer.result:
test case
mysql-test/t/group_by.test:
test case
mysql-test/t/join_outer.test:
test case
sql/sql_select.cc:
--remove wrong code
--use Field::real_maybe_null() check instead of
Field::maybe_null() and add addition check of
TABLE_LIST::outer_join
