libenchao commented on a change in pull request #11482: [FLINK-16581][table]
Minibatch deduplication lack state TTL bug fix
URL: https://github.com/apache/flink/pull/11482#discussion_r409348488
##########
File path:
flink-table/flink-table-runtime-blink/src/test/java/org/apache/flink/table/runtime/operators/deduplicate/MiniBatchDeduplicateKeepFirstRowFunctionTest.java
##########
@@ -71,4 +71,41 @@ public void testKeepFirstRowWithGenerateRetraction() throws
Exception {
testHarness.close();
}
+ @Test
+ public void tesKeepFirstRowWithStateTtl() throws Exception {
+ MiniBatchDeduplicateKeepFirstRowFunction func = new
MiniBatchDeduplicateKeepFirstRowFunction(typeSerializer,
minTime.toMilliseconds());
+ OneInputStreamOperatorTestHarness<BaseRow, BaseRow> testHarness
= createTestHarness(func);
+ testHarness.setup();
+ testHarness.open();
+ testHarness.processElement(record("book", 1L, 12));
+ testHarness.processElement(record("book", 2L, 11));
+ // output is empty because bundle not trigger yet.
+ Assert.assertTrue(testHarness.getOutput().isEmpty());
+ testHarness.processElement(record("book", 1L, 13));
+
+ testHarness.setStateTtlProcessingTime(30);
+ //Incremental cleanup is an eventual clean up, more state
access guarantee more expired state cleaned
+ for (long i = 3; i < 30; i++) {
+ testHarness.processElement(record("book", i, 20));
Review comment:
@Myasuka Thanks for the input.
+1 to use `NeverReturnExpired` strategy. We have got other questions from
user ML about this too. IMHO, `NeverReturnExpired` is more straight forward to
understand for users.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
