CTTY commented on code in PR #11947:
URL: https://github.com/apache/hudi/pull/11947#discussion_r1797513909
##########
hudi-utilities/src/main/java/org/apache/hudi/utilities/sources/HoodieIncrSource.java:
##########
@@ -175,33 +181,72 @@ public Pair<Option<Dataset<Row>>, String>
fetchNextBatch(Option<String> lastCkpt
IncrSourceHelper.MissingCheckpointStrategy missingCheckpointStrategy =
(containsConfigProperty(props,
HoodieIncrSourceConfig.MISSING_CHECKPOINT_STRATEGY))
? IncrSourceHelper.MissingCheckpointStrategy.valueOf(
- getStringWithAltKeys(props,
HoodieIncrSourceConfig.MISSING_CHECKPOINT_STRATEGY))
+ getStringWithAltKeys(props,
HoodieIncrSourceConfig.MISSING_CHECKPOINT_STRATEGY))
: null;
if (readLatestOnMissingCkpt) {
missingCheckpointStrategy =
IncrSourceHelper.MissingCheckpointStrategy.READ_LATEST;
}
- // Use begin Instant if set and non-empty
- Option<String> beginInstant =
- lastCkptStr.isPresent() ? lastCkptStr.get().isEmpty() ? Option.empty()
: lastCkptStr : Option.empty();
+ HoodieTableMetaClient metaClient = HoodieTableMetaClient.builder()
+
.setConf(HadoopFSUtils.getStorageConfWithCopy(sparkContext.hadoopConfiguration()))
+ .setBasePath(srcPath)
+ .setLoadActiveTimelineOnLoad(true)
+ .build();
+ int numInstantsFromConfig = getIntWithAltKeys(props,
HoodieIncrSourceConfig.NUM_INSTANTS_PER_FETCH);
+
+ String startTime;
+ if (lastCkptStr.isPresent() && !lastCkptStr.get().isEmpty()) {
+ startTime = lastCkptStr.get();
+ } else if (missingCheckpointStrategy != null) {
+ switch (missingCheckpointStrategy) {
+ case READ_UPTO_LATEST_COMMIT:
+ startTime = DEFAULT_BEGIN_TIMESTAMP;
+ // disrespect numInstantsFromConfig when reading up to latest
+ numInstantsFromConfig = -1;
+ break;
+ case READ_LATEST:
+ Option<HoodieInstant> lastInstant = metaClient
+ .getCommitsAndCompactionTimeline()
+ .filterCompletedInstants()
+ .lastInstant();
+ startTime = lastInstant
+ .map(hoodieInstant ->
getStrictlyLowerTimestamp(hoodieInstant.getCompletionTime()))
+ .orElse(DEFAULT_BEGIN_TIMESTAMP);
Review Comment:
We can not pass `null` or `Option.empty` directly to start time because it
will fall into this case:
https://github.com/apache/hudi/blob/master/hudi-common/src/main/java/org/apache/hudi/common/table/timeline/CompletionTimeQueryView.java#L278
if the `startTime` is `null` then it would disrespect the
`missingCheckpointStrategy` and will always read the latest snapshot i
--
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]
