lincoln-lil opened a new pull request, #20791:
URL: https://github.com/apache/flink/pull/20791
## What is the purpose of the change
Currently SinkUpsertMaterializer only update row by comparing the complete
row in anycase, but this may cause wrong result if input has upsertKey and also
non-deterministic column values, see such a case:
```java
@Test
public void testCdcWithNonDeterministicFuncSinkWithDifferentPk() {
tEnv.createTemporaryFunction(
"ndFunc", new JavaUserDefinedScalarFunctions.NonDeterministicUdf());
String cdcDdl =
"CREATE TABLE users (\n"
+ " user_id STRING,\n"
+ " user_name STRING,\n"
+ " email STRING,\n"
+ " balance DECIMAL(18,2),\n"
+ " primary key (user_id) not enforced\n"
+ ") WITH (\n"
+ " 'connector' = 'values',\n"
+ " 'changelog-mode' = 'I,UA,UB,D'\n"
+ ")";
String sinkTableDdl =
"CREATE TABLE sink (\n"
+ " user_id STRING,\n"
+ " user_name STRING,\n"
+ " email STRING,\n"
+ " balance DECIMAL(18,2),\n"
+ " PRIMARY KEY(email) NOT ENFORCED\n"
+ ") WITH(\n"
+ " 'connector' = 'values',\n"
+ " 'sink-insert-only' = 'false'\n"
+ ")";
tEnv.executeSql(cdcDdl);
tEnv.executeSql(sinkTableDdl);
util.verifyJsonPlan(
"insert into sink select user_id, ndFunc(user_name), email, balance from
users");
}
```
for original cdc source records:
```gittext
+I[user1, Tom, [email protected], 10.02],
-D[user1, Tom, [email protected], 10.02],
```
the above query cannot correctly delete the former insertion row because of
the non-deterministic column value 'ndFunc(user_name)'
this canbe solved by letting the SinkUpsertMaterializer be aware of input
upsertKey and update by it. This pr aims to address it.
## Brief change log
* Add projectRowType to RowTypeUtils and deprecate
AggCodeGenHelper#projectRowType
* Move non-deterministic test functions to userDefinedScalarFunctions
* Add upsertKey info for SinkUpsertMaterializer and do update via upsertKey
if has one (if no upsertKey, code path keeps unchanged)
## Verifying this change
add more related ut & it
## Does this pull request potentially affect one of the following parts:
- Dependencies (does it add or upgrade a dependency): (no)
- The public API, i.e., is any changed class annotated with
@Public(Evolving): (no)
- The serializers: (no )
- The runtime per-record code paths (performance sensitive): (no)
- Anything that affects deployment or recovery: JobManager (and its
components), Checkpointing, Kubernetes/Yarn, ZooKeeper: (no)
- The S3 file system connector: (no)
## Documentation
- Does this pull request introduce a new feature? (no)
--
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]
