matriv commented on a change in pull request #19232:
URL: https://github.com/apache/flink/pull/19232#discussion_r835143822
##########
File path:
flink-table/flink-table-planner/src/test/resources/org/apache/flink/table/planner/plan/nodes/exec/stream/GroupWindowAggregateJsonPlanTest_jsonplan/testProcTimeHopWindow.out
##########
@@ -338,6 +343,12 @@
}, {
"id" : 7,
"type" : "stream-exec-sink_1",
+ "configuration" : {
+ "table.exec.sink.keyed-shuffle" : "AUTO",
+ "table.exec.sink.not-null-enforcer" : "ERROR",
+ "table.exec.sink.type-length-enforcer" : "IGNORE",
Review comment:
https://issues.apache.org/jira/browse/FLINK-24753
https://issues.apache.org/jira/browse/FLINK-24413
We didn't have an actual discussion on the sink constraint enforcer. I would
argue that it's independent, and unless there is a Sink used that requires
string compliance for `CHAR/VARCHAR/BINARY/VARBINARY` (like an RDBMS), we
shouldn't enable it by default because of the performance overhead.
If users in their queries, use an explicit cast to those types, then by
default (legacyCastBehaviour=false) will get the padding & trimming. If they
use `CHAR/VARCHAR/BINARY/VARBINARY` as the target column type, without any
explicit casting, then, if the target sink doesn't have strong compliance,
everything will work out smoothly without performance penalty. If they need it
they need to enable it on demand.
WDYT?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
