danny0405 commented on a change in pull request #2961:
URL: https://github.com/apache/hudi/pull/2961#discussion_r635718807
##########
File path:
hudi-flink/src/test/java/org/apache/hudi/sink/TestWriteCopyOnWrite.java
##########
@@ -439,22 +439,22 @@ public void testInsertWithMiniBatches() throws Exception {
@Test
public void testInsertWithSmallBufferSize() throws Exception {
// reset the config option
- conf.setDouble(FlinkOptions.WRITE_TASK_MAX_SIZE, 200.001); // 1Kb buffer
size
+ conf.setDouble(FlinkOptions.WRITE_TASK_MAX_SIZE, 200.0006); // 630 bytes
buffer size
funcWrapper = new StreamWriteFunctionWrapper<>(tempFile.getAbsolutePath(),
conf);
// open the function and ingest data
funcWrapper.openFunction();
- // each record is 424 bytes. so 3 records expect to trigger buffer flush:
+ // each record is 208 bytes. so 4 records expect to trigger buffer flush:
// flush the max size bucket once at a time.
for (RowData rowData : TestData.DATA_SET_INSERT_DUPLICATES) {
funcWrapper.invoke(rowData);
}
Map<String, List<HoodieRecord>> dataBuffer = funcWrapper.getDataBuffer();
assertThat("Should have 1 data bucket", dataBuffer.size(), is(1));
- assertThat("2 records expect to flush out as a mini-batch",
+ assertThat("3 records expect to flush out as a mini-batch",
dataBuffer.values().stream().findFirst().map(List::size).orElse(-1),
- is(3));
+ is(2));
Review comment:
We do not flush the trigger record.
--
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:
[email protected]
