yihua commented on code in PR #11947:
URL: https://github.com/apache/hudi/pull/11947#discussion_r1802144497
##########
hudi-utilities/src/main/java/org/apache/hudi/utilities/sources/helpers/IncrSourceHelper.java:
##########
@@ -174,6 +170,64 @@ public static QueryInfo generateQueryInfo(JavaSparkContext
jssc, String srcBaseP
}
}
+ public static IncrementalQueryAnalyzer getIncrementalQueryAnalyzer(
+ JavaSparkContext jssc,
+ String srcPath,
+ Option<String> lastCkptStr,
+ MissingCheckpointStrategy missingCheckpointStrategy,
+ int numInstantsFromConfig,
+ Option<SourceProfile<Integer>> latestSourceProfile) {
+ HoodieTableMetaClient metaClient = HoodieTableMetaClient.builder()
+
.setConf(HadoopFSUtils.getStorageConfWithCopy(jssc.hadoopConfiguration()))
+ .setBasePath(srcPath)
+ .setLoadActiveTimelineOnLoad(true)
+ .build();
+
+ 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 ->
instantTimeMinusMillis(hoodieInstant.getCompletionTime(), 1))
Review Comment:
This is actually consistent with the logic in
`IncrSourceHelper#generateQueryInfo` where we treat begin checkpoint as
exclusive. The reason of doing this is to make sure the subsequent can be
applied. Another way is to special case this scenario by not going through the
incremental query analyzer and direct construct the `QueryContext` as the
instant to read is already known, i.e., the latest commit by completion time on
the timeline.
--
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]
