session: improvment refactor the logic of load data batch insert #11132
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov Report
@@ Coverage Diff @@
## master #11132 +/- ##
==============================================
+ Coverage 81.095% 81.104% +0.0089%
==============================================
Files 420 420
Lines 89622 89638 +16
==============================================
+ Hits 72679 72700 +21
+ Misses 11684 11670 -14
- Partials 5259 5268 +9 |
Codecov Report
@@ Coverage Diff @@
## master #11132 +/- ##
===========================================
Coverage 81.9464% 81.9464%
===========================================
Files 423 423
Lines 93239 93239
===========================================
Hits 76406 76406
Misses 11487 11487
Partials 5346 5346 |
crazycs520
reviewed
Jul 8, 2019
|
The title is a little long. Please shorten it. |
lysu
reviewed
Jul 9, 2019
executor/load_data.go
Outdated
| @@ -213,7 +215,7 @@ func (e *LoadDataInfo) getLine(prevData, curData []byte) ([]byte, []byte, bool) | |||
| // If it has the rest of data isn't completed the processing, then it returns without completed data. | |||
| // If the number of inserted rows reaches the batchRows, then the second return value is true. | |||
| // If prevData isn't nil and curData is nil, there are no other data to deal with and the isEOF is true. | |||
| func (e *LoadDataInfo) InsertData(prevData, curData []byte) ([]byte, bool, error) { | |||
| func (e *LoadDataInfo) InsertData(prevData, curData []byte, txnBatchCheckIns bool) ([]byte, bool, error) { | |||
There was a problem hiding this comment.
maybe better add some comment for txnBatchCheckIns argument
executor/load_data.go
Outdated
| if err != nil { | ||
| return nil, reachLimit, err | ||
| if !txnBatchCheckIns { | ||
| err := e.CheckAndInsertOneBatch() |
There was a problem hiding this comment.
after this PR, it seems only unit-test logic will step into here, how about this code snippet and remove txnBatchCheckIns let it always be true and modify unit-test to test real check logic
cfzjywxk
changed the title
…en once per transaction
cfzjywxk
changed the title
|
/run-all-tests |
1 similar comment
|
/run-all-tests |
lysu
added
component/server
type/enhancement
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
…ce per transaction
What problem does this PR solve?
issue 3092
load data batch checked all rows for one packet instead of one transaction, will cause too many batch-get to the storage node.
We could do one batch-get only for one transaction.
local rough tests show that
batch get per transaction may not be always better,
the cost depends on (batchget efficiency with more keys ) and batchget call rt
generally, slower batchget calls result in more load data time cost
What is changed and how it works?
make batchget happens one per transcation but not per packet
Check List
Tests
Code changes
Side effects
Related changes