mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-16 03:52:35 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/mysql-test/main/index_intersect.test
Monty eb483c5181 Updated optimizer costs in multi_range_read_info_const() and sql_select.cc 
- multi_range_read_info_const now uses the new records_in_range interface
- Added handler::avg_io_cost()
- Don't calculate avg_io_cost() in get_sweep_read_cost if avg_io_cost is
  not 1.0.  In this case we trust the avg_io_cost() from the handler.
- Changed test_quick_select to use TIME_FOR_COMPARE instead of
  TIME_FOR_COMPARE_IDX to align this with the rest of the code.
- Fixed bug when using test_if_cheaper_ordering where we didn't use
  keyread if index was changed
- Fixed a bug where we didn't use index only read when using order-by-index
- Added keyread_time() to HEAP.
  The default keyread_time() was optimized for blocks and not suitable for
  HEAP. The effect was the HEAP prefered table scans over ranges for btree
  indexes.
- Fixed get_sweep_read_cost() for HEAP tables
- Ensure that range and ref have same cost for simple ranges
  Added a small cost (MULTI_RANGE_READ_SETUP_COST) to ranges to ensure
  we favior ref for range for simple queries.
- Fixed that matching_candidates_in_table() uses same number of records
  as the rest of the optimizer
- Added avg_io_cost() to JT_EQ_REF cost. This helps calculate the cost for
  HEAP and temporary tables better. A few tests changed because of this.
- heap::read_time() and heap::keyread_time() adjusted to not add +1.
  This was to ensure that handler::keyread_time() doesn't give
  higher cost for heap tables than for normal tables. One effect of
  this is that heap and derived tables stored in heap will prefer
  key access as this is now regarded as cheap.
- Changed cost for index read in sql_select.cc to match
  multi_range_read_info_const(). All index cost calculation is now
  done trough one function.
- 'ref' will now use quick_cost for keys if it exists. This is done
  so that for '=' ranges, 'ref' is prefered over 'range'.
- scan_time() now takes avg_io_costs() into account
- get_delayed_table_estimates() uses block_size and avg_io_cost()
- Removed default argument to test_if_order_by_key(); simplifies code
2020-03-27 03:58:32 +02:00

479 lines
14 KiB
Text
Raw Blame History

--source include/default_optimizer_switch.inc
--source include/default_charset.inc
--disable_warnings
DROP TABLE IF EXISTS t1,t2,t3,t4;
DROP DATABASE IF EXISTS world;
--enable_warnings
set names utf8;
CREATE DATABASE world;
use world;
--source include/world_schema.inc
--disable_query_log
--disable_result_log
--disable_warnings
--source include/world.inc
--enable_warnings
--enable_result_log
--enable_query_log
SELECT COUNT(*) FROM Country;
SELECT COUNT(*) FROM City;
SELECT COUNT(*) FROM CountryLanguage;
CREATE INDEX Name ON City(Name);
--disable_query_log
--disable_result_log
--disable_warnings
ANALYZE TABLE City;
--enable_warnings
--enable_result_log
--enable_query_log
SET SESSION optimizer_switch='rowid_filter=off';
SET SESSION optimizer_switch='index_merge_sort_intersection=on';
SELECT COUNT(*) FROM City;
# The output of the next 6 queries tells us about selectivities
# of the conditions utilized in 4 queries following after them
SELECT COUNT(*) FROM City WHERE Name LIKE 'C%';
SELECT COUNT(*) FROM City WHERE Name LIKE 'M%';
SELECT COUNT(*) FROM City WHERE Population > 1000000;
SELECT COUNT(*) FROM City WHERE Population > 1500000;
SELECT COUNT(*) FROM City WHERE Population > 300000;
SELECT COUNT(*) FROM City WHERE Population > 7000000;
# The pattern of the WHERE condition used in the following 4 queries is
# range(key1) AND range(key2)
# Varying values of the constants in the conjuncts of the condition
# we can get either an index intersection retrieval over key1 and key2
# or a range index scan for one of these indexes
--replace_column 9 #
EXPLAIN
SELECT * FROM City WHERE
Name LIKE 'C%' AND Population > 1000000;
--replace_column 9 #
EXPLAIN
SELECT * FROM City WHERE
Name LIKE 'M%' AND Population > 1500000;
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Name LIKE 'M%' AND Population > 300000;
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Name LIKE 'M%' AND Population > 7000000;
# The following 8 queries check that
# the previous 4 plans are valid and return
# the correct results when executed
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE Name LIKE 'C%' AND Population > 1000000;
--sorted_result
SELECT * FROM City
WHERE Name LIKE 'C%' AND Population > 1000000;
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE Name LIKE 'M%' AND Population > 1500000;
--sorted_result
SELECT * FROM City
WHERE Name LIKE 'M%' AND Population > 1500000;
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE Name LIKE 'M%' AND Population > 300000;
--sorted_result
SELECT * FROM City
WHERE Name LIKE 'M%' AND Population > 300000;
SELECT * FROM City USE INDEX ()
WHERE Name LIKE 'M%' AND Population > 7000000;
SELECT * FROM City
WHERE Name LIKE 'M%' AND Population > 7000000;
# The output of the next 7 queries tells us about selectivities
# of the conditions utilized in 3 queries following after them
SELECT COUNT(*) FROM City WHERE Name BETWEEN 'M' AND 'N';
SELECT COUNT(*) FROM City WHERE Name BETWEEN 'G' AND 'J';
SELECT COUNT(*) FROM City WHERE Name BETWEEN 'G' AND 'K';
SELECT COUNT(*) FROM City WHERE Population > 1000000;
SELECT COUNT(*) FROM City WHERE Population > 500000;
SELECT COUNT(*) FROM City WHERE Country LIKE 'C%';
SELECT COUNT(*) FROM City WHERE Country LIKE 'B%';
SELECT COUNT(*) FROM City WHERE Country LIKE 'J%';
# The pattern of the WHERE condition used in the following 3 queries is
# range(key1) AND range(key2) AND range(key3)
# Varying values of the constants in the conjuncts of the condition
# we can get index intersection over different pairs of keys:
# over(key1,key2), over(key1,key3) and over(key2,key3)
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Name BETWEEN 'M' AND 'N' AND Population > 1000000 AND Country LIKE 'C%';
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'K' AND Population > 1000000 AND Country LIKE 'J%';
--replace_column 7 # 9 #
--replace_result Population,Country,Name Population,Name,Country
EXPLAIN
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'K' AND Population > 500000 AND Country LIKE 'C%';
# The following 6 queries check that
# the previous 3 plans are valid and return
# the correct results when executed
SELECT * FROM City USE INDEX ()
WHERE Name BETWEEN 'M' AND 'N' AND Population > 1000000 AND Country LIKE 'C%';
SELECT * FROM City
WHERE Name BETWEEN 'M' AND 'N' AND Population > 1000000 AND Country LIKE 'C%';
SELECT * FROM City USE INDEX ()
WHERE Name BETWEEN 'G' AND 'K' AND Population > 1000000 AND Country LIKE 'J%';
--sorted_result
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'K' AND Population > 1000000 AND Country LIKE 'J%';
SELECT * FROM City USE INDEX ()
WHERE Name BETWEEN 'G' AND 'K' AND Population > 500000 AND Country LIKE 'C%';
--sorted_result
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'K' AND Population > 500000 AND Country LIKE 'C%';
# The output of the next 12 queries tells us about selectivities
# of the conditions utilized in 5 queries following after them
SELECT COUNT(*) FROM City WHERE ID BETWEEN 501 AND 1000;
SELECT COUNT(*) FROM City WHERE ID BETWEEN 1 AND 500;
SELECT COUNT(*) FROM City WHERE ID BETWEEN 2001 AND 2500;
SELECT COUNT(*) FROM City WHERE ID BETWEEN 3701 AND 4000;
SELECT COUNT(*) FROM City WHERE Population > 700000;
SELECT COUNT(*) FROM City WHERE Population > 1000000;
SELECT COUNT(*) FROM City WHERE Population > 300000;
SELECT COUNT(*) FROM City WHERE Population > 600000;
SELECT COUNT(*) FROM City WHERE Country LIKE 'C%';
SELECT COUNT(*) FROM City WHERE Country LIKE 'A%';
SELECT COUNT(*) FROM City WHERE Country LIKE 'H%';
SELECT COUNT(*) FROM City WHERE Country BETWEEN 'S' AND 'Z';
# The pattern of the WHERE condition used in the following 5 queries is
# range(key1) AND range(key2) AND range(key3)
# with key1 happens to be a primary key (it matters only for InnoDB)
# Varying values of the constants in the conjuncts of the condition
# we can get index intersection either over all three keys, or over
# different pairs, or a range scan over one of these keys.
# Bear in mind that the condition (Country LIKE 'A%') is actually
# equivalent to the condition (Country BETWEEN 'A' AND 'B') for the
# tested instance the table City.
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE ID BETWEEN 501 AND 1000 AND Population > 700000 AND Country LIKE 'C%';
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE ID BETWEEN 1 AND 500 AND Population > 700000 AND Country LIKE 'C%';
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE ID BETWEEN 2001 AND 2500 AND Population > 300000 AND Country LIKE 'H%';
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE ID BETWEEN 3701 AND 4000 AND Population > 1000000
AND Country BETWEEN 'S' AND 'Z';
--replace_column 9 #
--replace_result PRIMARY,Country,Population PRIMARY,Population,Country 4,7,4 4,4,7
EXPLAIN
SELECT * FROM City
WHERE ID BETWEEN 3001 AND 4000 AND Population > 600000
AND Country BETWEEN 'S' AND 'Z' ;
# The following 10 queries check that
# the previous 5 plans are valid and return
# the correct results when executed
SELECT * FROM City USE INDEX ()
WHERE ID BETWEEN 501 AND 1000 AND Population > 700000 AND Country LIKE 'C%';
SELECT * FROM City
WHERE ID BETWEEN 501 AND 1000 AND Population > 700000 AND Country LIKE 'C%';
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE ID BETWEEN 1 AND 500 AND Population > 700000 AND Country LIKE 'C%';
--sorted_result
SELECT * FROM City
WHERE ID BETWEEN 1 AND 500 AND Population > 700000 AND Country LIKE 'C%';
SELECT * FROM City USE INDEX ()
WHERE ID BETWEEN 2001 AND 2500 AND Population > 300000 AND Country LIKE 'H%';
SELECT * FROM City
WHERE ID BETWEEN 2001 AND 2500 AND Population > 300000 AND Country LIKE 'H%';
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE ID BETWEEN 3701 AND 4000 AND Population > 700000
AND Country BETWEEN 'S' AND 'Z';
--sorted_result
SELECT * FROM City
WHERE ID BETWEEN 3701 AND 4000 AND Population > 700000
AND Country BETWEEN 'S' AND 'Z';
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE ID BETWEEN 3001 AND 4000 AND Population > 600000
AND Country BETWEEN 'S' AND 'Z' ;
--sorted_result
SELECT * FROM City
WHERE ID BETWEEN 3001 AND 4000 AND Population > 600000
AND Country BETWEEN 'S' AND 'Z' ;
# Originally this was just sort_buffer_size=2048. Then, it started
# failing on 32bit due to different cost number in
# Unique::get_use_cost() because of sizeof(sizeof(TREE_ELEMENT)+key_size)
# On 64 bit: Unique object element_size=32, which gives 2048/32= 64 elements
# in the tree.
# On 32 bit: Unique object element_size=24.
# If we want 64 elements in the tree, we need 64*24=1536 as sort_buffer_size.
# The purpose of setting sort_buffer_size is to show that some of the following
# explains should use 'index_merge' while others should use range
# If the following code causes future problems, the other option would be
# to create a separate result-.diff file for 32 bit.
SET SESSION sort_buffer_size = IF(@@version_compile_machine like '%64%', 2048, 1536);
# The following EXPLAIN command demonstrate that the execution plans
# may be different if sort_buffer_size is set to a small value
--replace_column 9 #
EXPLAIN
SELECT * FROM City WHERE
Name LIKE 'C%' AND Population > 1000000;
--replace_column 9 #
EXPLAIN
SELECT * FROM City WHERE
Name LIKE 'M%' AND Population > 1500000;
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'K' AND Population > 1000000 AND Country LIKE 'J%';
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'J' AND Population > 500000 AND Country LIKE 'C%';
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE ID BETWEEN 1 AND 500 AND Population > 700000 AND Country LIKE 'C%';
--replace_column 9 #
--replace_result PRIMARY,Country,Population PRIMARY,Population,Country 4,7,4 4,4,7
EXPLAIN
SELECT * FROM City
WHERE ID BETWEEN 3001 AND 4000 AND Population > 600000
AND Country BETWEEN 'S' AND 'Z';
#Yet the query themselves return the correct results in this case as well
--sorted_result
SELECT * FROM City WHERE
Name LIKE 'C%' AND Population > 1000000;
--sorted_result
SELECT * FROM City WHERE
Name LIKE 'M%' AND Population > 1500000;
--sorted_result
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'J' AND Population > 700000 AND Country LIKE 'J%';
--sorted_result
SELECT * FROM City
WHERE Name BETWEEN 'G' AND 'J' AND Population > 500000 AND Country LIKE 'C%';
SELECT * FROM City
WHERE ID BETWEEN 1 AND 500 AND Population > 700000 AND Country LIKE 'C%';
--sorted_result
SELECT * FROM City
WHERE ID BETWEEN 3001 AND 4000 AND Population > 600000
AND Country BETWEEN 'S' AND 'Z';
SET SESSION sort_buffer_size = default;
# Instead of the index on the column Country create two compound indexes
# including this column as the first component
DROP INDEX Country ON City;
CREATE INDEX CountryID ON City(Country,ID);
CREATE INDEX CountryName ON City(Country,Name);
--disable_query_log
--disable_result_log
--disable_warnings
ANALYZE TABLE City;
--enable_warnings
--enable_result_log
--enable_query_log
# Check that the first component of a compound index can be used for
# index intersection, even in the cases when we have a ref access
# for this component
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Country LIKE 'M%' AND Population > 1000000;
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Country='USA' AND Population > 1000000;
--replace_column 9 #
EXPLAIN
SELECT * FROM City
WHERE Country='USA' AND Population > 1500000 AND Name LIKE 'C%';
# Check that the previous 3 plans return the right results when executed
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE Country LIKE 'M%' AND Population > 1000000;
--sorted_result
SELECT * FROM City
WHERE Country LIKE 'M%' AND Population > 1000000;
--sorted_result
SELECT * FROM City USE INDEX ()
WHERE Country='USA' AND Population > 1000000;
--sorted_result
SELECT * FROM City
WHERE Country='USA' AND Population > 1000000;
SELECT * FROM City USE INDEX ()
WHERE Country='USA' AND Population > 1500000 AND Name LIKE 'C%';
--sorted_result
SELECT * FROM City
WHERE Country='USA' AND Population > 1500000 AND Name LIKE 'C%';
#
# Bug #754521: wrong cost of index intersection leading
# to the choice of a suboptimal execution plan
#
--replace_column 9 #
EXPLAIN
SELECT * FROM City, Country
WHERE City.Name LIKE 'C%' AND City.Population > 1000000 AND
Country.Code=City.Country;
DROP DATABASE world;
use test;
#
# Bug #684086: crash with EXPLAIN in InnoDB for index intersection
# of two indexes one of which is primary
#
CREATE TABLE t1 (
f1 int,
f4 varchar(32),
f5 int,
PRIMARY KEY (f1),
KEY (f4)
);
INSERT INTO t1 VALUES
(5,'H',1), (9,'g',0), (527,'i',0), (528,'y',1), (529,'S',6),
(530,'m',7), (531,'b',2), (532,'N',1), (533,'V',NULL), (534,'l',1),
(535,'M',0), (536,'w',1), (537,'j',5), (538,'l',0), (539,'n',2),
(540,'m',2), (541,'r',2), (542,'l',2), (543,'h',3),(544,'o',0),
(956,'h',0), (957,'g',0), (958,'W',5), (959,'s',3), (960,'w',0),
(961,'q',0), (962,'e',NULL), (963,'u',7), (964,'q',1), (965,'N',NULL),
(966,'e',0), (967,'t',3), (968,'e',6), (969,'f',NULL), (970,'j',0),
(971,'s',3), (972,'I',0), (973,'h',4), (974,'g',1), (975,'s',0),
(976,'r',3), (977,'x',1), (978,'v',8), (979,'j',NULL), (980,'z',7),
(981,'t',9), (982,'j',5), (983,'u',NULL), (984,'g',6), (985,'w',1),
(986,'h',1), (987,'v',0), (988,'v',0), (989,'c',2), (990,'b',7),
(991,'z',0), (992,'M',1), (993,'u',2), (994,'r',2), (995,'b',4),
(996,'A',2), (997,'u',0), (998,'a',0), (999,'j',2), (1,'I',2);
--replace_column 9 #
EXPLAIN
SELECT * FROM t1
WHERE (f1 < 535 OR f1 > 985) AND ( f4='r' OR f4 LIKE 'a%' ) ;
--sorted_result
SELECT * FROM t1
WHERE (f1 < 535 OR f1 > 985) AND ( f4='r' OR f4 LIKE 'a%' ) ;
DROP TABLE t1;
SET SESSION optimizer_switch='index_merge_sort_intersection=on';
SET SESSION optimizer_switch='rowid_filter=default';
Reference in a new issue View git blame Copy permalink