2550969cae
aggregation, plan: split the presentation and evaluation layers of aggregation functions ( #5635 )
2018-01-15 15:48:03 +08:00
ca97bfffe0
tables: when importing data, table.AddRecord has no needs to use temp BufferStore. ( #5617 )
2018-01-15 15:13:23 +08:00
d568bf7f7c
executor: refactor the way to compare rows in MergeJoinExec ( #5640 )
2018-01-15 12:41:07 +08:00
4d50ae22bc
kvencode: enable prepare plan cache. ( #5627 )
2018-01-14 15:04:36 +08:00
3ce567630b
executor: store column's fieldType in executor ( #5619 )
2018-01-12 16:57:31 +08:00
3cff21e5b1
plan, executor: avoid children type assertion. ( #5616 )
2018-01-12 15:39:50 +08:00
18ba3470f6
stats: adjust CM sketch default size ( #5614 )
2018-01-12 14:29:46 +08:00
31b606d30d
executer: optimization for insert ignore ( #5508 )
2018-01-12 11:09:25 +08:00
4dd1fc208d
codec, expression: support encoding a Chunk row ( #5578 )
2018-01-11 20:29:31 +08:00
f8a2c807fd
executor: support Chunk for ExecuteExec ( #5410 )
2018-01-11 18:38:12 +08:00
5a555703c0
*: garbage collect useless stats info ( #5587 )
2018-01-11 17:10:30 +08:00
20f85865f0
executor: close hashjoin goroutines as soon as possible to avoid unexpected error ( #5620 )
2018-01-11 16:22:44 +08:00
f4fa60d306
executor, expression: support chunk for HashJoinExec ( #5439 )
2018-01-10 22:06:58 +08:00
6f2a8577b9
chunk: "setMutRowJSON" should allocate one more byte for null bitmap ( #5603 )
...
* chunk: "setMutRowJSON" should allocate one more byte for null bitmap
* address comment
2018-01-10 21:35:01 +08:00
3e298e3390
executor: "fetchShowTableStatus" should append an integer to the third column instead of a string ( #5608 )
2018-01-10 17:05:23 +08:00
3cd10669a0
plan, executor: change 'out of order' to 'keep order'. ( #5607 )
2018-01-10 15:57:31 +08:00
4136b81c23
*: merge IntColumnRange with NewRange. ( #5591 )
2018-01-10 15:50:58 +08:00
b7e7cd5d9a
executor: lazily initialize InsertValues.colDefaultVals, because not every insert statement needs it. ( #5589 )
2018-01-09 22:36:04 +08:00
e9b450aa0a
table: InsertExec eliminate unnecessary CastValue. ( #5581 )
2018-01-08 19:21:04 +08:00
4c63b26270
plan: fix issue#5570. ( #5574 )
2018-01-08 16:41:03 +08:00
eddff3429a
*: importing data reduce memory allocations. ( #5549 )
2018-01-07 12:48:07 +08:00
ecbd60c3ff
*: add arguement StatementContext to Encode functions. ( #5566 )
2018-01-06 17:31:08 +08:00
de992eb570
expression: only do a shallow copy when evaluating a "Column" expression ( #5542 )
2018-01-06 12:24:30 +08:00
bc459c5261
executor: support Chunk for HashAggExec ( #5244 )
2018-01-06 10:21:52 +08:00
c45111d9f6
stats: use more compact structure for histograms ( #5539 )
...
* stats: use more compact structure for histograms
* address comments
* address comments
* refine code
* address comment
2018-01-05 20:43:01 +08:00
8b149c3323
Prealloc / collides with package name ( #5560 )
2018-01-05 19:15:54 +08:00
73e47e0e30
plan: refine the design of schema ( #5541 )
2018-01-05 12:54:24 +08:00
c19307c41c
util, executor: refactor Chunk.AppendRow to handle virtual row ( #5563 )
2018-01-05 12:35:21 +08:00
04ef7d7992
*: shard implicit row ID ( #5513 )
2018-01-04 16:32:41 +08:00
019dbb7c9b
executor: support Chunk for StreamAggExec ( #5490 )
2018-01-03 19:47:18 +08:00
b10a65dca1
plan: covert max/min to topN. ( #5516 )
2018-01-02 22:30:17 +08:00
36d9b61365
*: AddIndex skip checking unique constraint when TiDB is importing data. ( #5535 )
2018-01-02 20:08:49 +08:00
93b52805c3
ast, executor: set Fields for SelectStmt in PrepareExec ( #5504 )
2017-12-27 17:16:49 +08:00
b0921f9d27
executor: implement NextChunk interface for UnionScanExec ( #5432 )
2017-12-26 20:08:11 +08:00
b543bcc7df
executor: joinGenerator use chunk.Iterator ( #5500 )
...
So NestedLoopApply can use *chunk.List to generate results and reuse the memory.
2017-12-26 18:47:29 +08:00
140c3d0739
executor: "InsertExec.NextChunk" should use function "getRowsSelectChunk" ( #5493 )
2017-12-26 18:24:41 +08:00
95a5c8ce5a
expresion, executor: correct the type inference of function "sum" and "avg" ( #5495 )
2017-12-26 18:16:52 +08:00
fa62e9fd01
*: merge ranger.IndexRange and ranger.ColumnRange. ( #5485 )
2017-12-26 16:08:36 +08:00
ccc10a56a6
executor: support Chunk for NestedLoopApply ( #5475 )
2017-12-26 13:37:19 +08:00
46fd1eb536
executor: support Chunk for ShowExec ( #5456 )
2017-12-25 21:29:56 +08:00
a7beb99d75
*: support pushing down stream aggregation on mocktikv ( #5444 )
2017-12-25 15:40:53 +08:00
cc0c09820d
executor, plan: reading from memtable should only take the needed columns ( #5481 )
2017-12-25 15:15:18 +08:00
48aae7b6ef
*: support show auto increment ID ( #5470 )
2017-12-22 20:16:11 +08:00
7159dd0ef6
executor: support Chunk for TableScanExec ( #5443 )
2017-12-22 19:52:43 +08:00
232ecbeca0
executor: support Chunk for MergeJoinExec ( #5312 )
2017-12-22 16:47:22 +08:00
855df184e1
executor: support Chunk for TableDualExec ( #5395 )
2017-12-22 16:37:21 +08:00
a386528393
executor: merge ApplyExec and NestedLoopJoin into NestedLoopApply. ( #5471 )
...
NestedLoopJoin is only used in ApplyJoinExec and ApplyJoinExec only has NestedLoopJoin.
Merge them into one simplifies the logic.
2017-12-22 13:44:43 +08:00
ec1b40d5b3
executor: Close() executors when Open() returns an error to avoid goroutine leak ( #5469 )
2017-12-22 12:38:07 +08:00
2c6d1cf8fa
plan,executor: remove HashSemiJoin ( #5467 )
...
HashSemiJoin is only used in Apply, it exists only because NestedLoopJoin doesn't support SemiJoin type.
This PR support SemiJoin type in NestedLoopJoin executor, so HashSemiJoin can be removed.
2017-12-22 12:25:42 +08:00
90401445d2
*: collect and store query feedback ( #5438 )
2017-12-22 11:33:05 +08:00
