-
Notifications
You must be signed in to change notification settings - Fork 101
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
SNOW-824475 Support Spark 3.4 #510
Conversation
Codecov Report
@@ Coverage Diff @@
## master #510 +/- ##
==========================================
- Coverage 89.59% 89.52% -0.07%
==========================================
Files 52 52
Lines 4487 4487
Branches 744 741 -3
==========================================
- Hits 4020 4017 -3
- Misses 467 470 +3 |
400c86f
to
c1d7b56
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
|( SELECT * FROM ( SELECT * FROM ( $test_table2 ) AS | ||
|"SF_CONNECTOR_QUERY_ALIAS" ) AS "SUBQUERY_0" WHERE | ||
|( "SUBQUERY_0"."O" IS NOT NULL ) ) AS "SUBQUERY_1" | ||
""".stripMargin |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If I understand correctly, we issue two SQLs now instead of one for this push down? Is that a performance issue?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is test-only and it doesn't cause any perf problem for production.
- The generated query for different spark may be different.
- To make the test case to work for different Spark versions, the
expectedMultiplyQueries
includes the query for spark 3.4 and previous version. If the query matches either one, the test case will succeed.
SNOW-824475 Support Spark 3.4
The main changes for Spark 3.4 affects Spark Connector are:
Cast
ansiEnabled: Boolean
evalMode: EvalMode.Value
LEGACY
,ANSI
,TRY
LEGACY
. It will make SC to be in consistent behaviour.JoinHint
argument forScalarSubquery
.PromotePrecision
is removed from Spark 3.4ansiEnabled
forRound
.