77572f873d
planner: do not push down the aggregation function with correlated column ( #21453 )
2020-12-08 10:54:24 +08:00
e33b37789c
planner: add projection if expand virtual generated column ( #20478 )
...
Signed-off-by: wjhuang2016 <huangwenjun1997@gmail.com >
2020-12-03 21:42:49 +08:00
76c1d28af9
planner: support complex join in broadcast join ( #20196 )
2020-11-05 16:29:35 +08:00
aa2b83210d
planner: make pushdown limit use its child schema ( #20807 )
2020-11-04 15:58:34 +08:00
369c12f779
[planner, executor, kv]: refine and extract mpp codes to the right place ( #20691 )
2020-11-02 14:30:47 +08:00
b1b1da22e5
planner: fix wrong argument for split aggregation ( #20474 )
2020-10-29 15:08:42 +08:00
e52cb3714e
planner: Do not pushdown unsupported function to tiflash in TopN operator ( #19363 )
...
Co-authored-by: ti-srebot <66930949+ti-srebot@users.noreply.github.com >
2020-08-25 17:20:18 +08:00
977449f8b0
planner: fix a bug that can't find column in indexLookupJoin with virtual generated column ( #18516 )
2020-08-25 16:09:31 +08:00
3ac0baf61c
*: fix unable prune problem caused by mismatch column infos ( #19391 )
2020-08-25 12:24:48 +08:00
4e829aaee7
planner,executor: fix index merge for partition table ( #18940 )
...
Co-authored-by: ti-srebot <66930949+ti-srebot@users.noreply.github.com >
2020-08-04 20:38:17 +08:00
5dd7cd8553
planner/core: implement Explain for the new partition table ( #18834 )
...
Co-authored-by: lysu <sulifx@gmail.com >
2020-07-30 15:37:02 +08:00
a1dc4381c5
planner,executor: implement PartitionTable executor for the three reader ( #18823 )
2020-07-30 11:00:55 +08:00
29178dfa8d
planner, executor: support broadcast join for tiflash engine. ( #17232 )
...
* support batch cop for tiflash
* support batch cop
* support join push down to tiflash
* refine
* change pb
* push join
* fix
* add hint
* refine hint
* add ranges
* fix
* fix
* fix push down
* fix index
* enable distsql for join
* add a session var to disable/enable broadcast join
* fix bug
* fix bug
* tiny fix
* enable cast decimal pushdown to tiflash
* fix
* fix bc join bug
* make broadcast plan stable
* refine code
* fix bug
* basic support for multi table broadcast join
* fix bug
* basic cbo for broadcast join
* improve
* fix bug
* remote useless code
* add tests
* pass unit tests
* refine code
* support execute summary info for broadcast join
* fix bug in explain for broadcast join
* format code
* remove un-needed code
* fix make dev
* address comments
* Hanfei/join merge (#7 )
* enable exec details for batch cop
* format code
* fix test
* change tidb_opt_broadcast_join to global vars
* Ban cartesian join to be pushed down to TiFlash (#8 )
* merge master (#10 )
* merge master
* fix bug
* fix bug
* fix ut
* check session var conflict
* Add perfer local hint for broadcast join (#12 )
* update
* remove useless code
* remove useless code
* update parser
* add test for prefer local join
* use bcj_local
* update go.mod
* refine planner
* refine comments
* fix make dev
* fix make dev
* update parser
* address comments
* fix make dev
* disable broadcast join when new collation is enabled
* Update planner/core/exhaust_physical_plans.go
Co-authored-by: Zhuomin(Charming) Liu <lzmhhh123@gmail.com >
* address comments
* fix
* address comments
* fix tests
* address comments
Co-authored-by: xufei <xufeixw@mail.ustc.edu.cn >
Co-authored-by: xufei <xufei@pingcap.com >
Co-authored-by: 虎 <ichneumon.hu@foxmail.com >
Co-authored-by: Zhuomin(Charming) Liu <lzmhhh123@gmail.com >
2020-07-27 12:41:36 +08:00
a6971dd685
*: support 'admin check table' on clustered index table ( #18346 )
2020-07-14 15:59:37 +08:00
daad2699dd
*: support clustered index for index join ( #18204 )
2020-07-01 16:15:51 +08:00
5250f7f53f
*: support clustered index in double read ( #18127 )
2020-06-23 15:33:38 +08:00
c98514e751
executor: unify executor concurrency ( #16999 )
2020-06-22 17:39:32 +08:00
978370f7cb
executor: add new agg function APPROX_COUNT_DISTINCT ( #17175 )
2020-06-18 22:00:14 +08:00
7ebcc20398
executor: support GROUP_CONCAT(ORDER BY) ( #16591 )
2020-05-06 18:41:22 +08:00
8f000fc353
planner: push avg & distinct functions across union ( #16344 )
2020-04-15 18:07:37 +08:00
49ee75a0a5
planner, executor: enable inline projection for merge join ( #15463 )
2020-04-09 15:58:32 +08:00
77f970c48c
planner: fix push down distinct when need to inject projection ( #15997 )
2020-04-02 11:43:01 +08:00
4eb9ca3d3f
planner: push aggregation functions with distinct to cop ( #15500 )
2020-04-01 20:49:14 +08:00
5282d089c7
executor: support generated column on point get and batch point get ( #15544 )
2020-03-23 14:21:35 +08:00
a999ef6981
expression: support different expr push down for TiKV and TiFlash ( #15174 )
2020-03-11 16:16:03 +08:00
9e98f706ae
planner, executor: support point get as a option of DataSource ( #14775 )
2020-03-09 10:47:16 +08:00
1703ee82cd
planner: not to generate partial agg when cop task has root conditions ( #15112 )
2020-03-04 19:57:17 +08:00
c1114b9724
planner, executor: enable inline projection for index joins ( #15004 )
2020-03-02 15:32:09 +08:00
6d00e2169b
planner: enable inline projection for hash join ( #14783 )
2020-02-28 16:28:10 +08:00
b64ad65636
planner: improve projection to keep order or keep index in some cases ( #14510 )
2020-02-26 14:02:36 +08:00
3399f4be31
planner: distinguish FullScan and the RangeScan in explain/desc info ( #14696 )
2020-02-24 15:23:06 +08:00
6e867b7cbc
planner: consider disk cost in Sort ( #14708 )
2020-02-12 11:40:47 +08:00
e39b504177
planner: adjust the cost model of Apply ( #13550 )
2020-01-15 19:43:50 +08:00
08be35d778
planner: fix wrong stats for physical union scan ( #14093 )
2019-12-19 10:37:16 +08:00
02408dbd04
planner: support cost model for Chunk encode format. ( #13461 )
2019-12-18 15:47:16 +08:00
9b7b2c0d50
planner/core: fix bug when using generated column in aggregate statement ( #14111 )
2019-12-18 14:46:47 +08:00
438d39ebcd
*: revert #10124 and remove sql mode PadCharToFullLength ( #14007 )
2019-12-17 15:25:06 +08:00
71e1288f00
planner/cascades: support implementation rule Apply and MaxOneRow. ( #13873 )
2019-12-06 17:19:48 +08:00
5b6422df58
*: support read remote TiDB memory table of cluster ( #13065 )
2019-12-04 13:05:36 +08:00
5a589c9994
planner: consider disk cost in hashJoin ( #13246 )
2019-12-03 13:43:19 +08:00
96069cd749
planner: reuse GroupBy Columns when build finalModeAgg ( #13653 )
2019-11-21 13:53:37 +08:00
02916a6778
planner/cascades: add transformation PushSelDownJoin & Implementation for HashJoin ( #13470 )
2019-11-20 16:13:13 +08:00
161a2eb454
planner/cascades: add transformation rule PushAggDownGather ( #13049 )
2019-11-15 18:09:35 +08:00
aca3f47292
*: Refactoring reading logic of virtual generate column ( #12407 )
2019-11-13 17:39:53 +08:00
4dcd7d7938
planner: consider build hashtable using the outer side for outer join during plan building ( #12883 )
2019-11-08 17:43:53 +08:00
6fd74f3e4c
planner: support cost model for tiflash table scan ( #12868 )
2019-10-29 16:05:51 +08:00
36f0f37b40
expression, planner: remove some fields from `expression.Colum… ( #12573 )
2019-10-29 15:48:26 +08:00
b8cd65714f
*: record and print the plan in slow log. ( #12179 )
2019-10-17 22:13:19 +08:00
45bc789a0f
planner: stream agg should not be pushed to double read ( #12443 )
...
For the following two reason, we should not push stream agg down to double read
- The aggregate will lost the handle information
- There's no sort operator. The second read is ordered with pk, not by index.
2019-10-12 19:22:04 +08:00
7ffa4500e6
planner: support a hint to read from tiflash in planner ( #12479 )
2019-10-11 16:36:35 +08:00
