[
https://issues.apache.org/jira/browse/BEAM-14502?focusedWorklogId=775567&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-775567
]
ASF GitHub Bot logged work on BEAM-14502:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 27/May/22 19:54
Start Date: 27/May/22 19:54
Worklog Time Spent: 10m
Work Description: diegomez17 commented on code in PR #16939:
URL: https://github.com/apache/beam/pull/16939#discussion_r883942676
##########
sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigtable/BigtableIO.java:
##########
@@ -402,6 +404,28 @@ public Read withRowFilter(RowFilter filter) {
return withRowFilter(StaticValueProvider.of(filter));
}
+ /**
+ * Returns a new {@link BigtableIO.Read} that will break up read requests
into smaller batches.
+ * This function will switch the base BigtableIO.Reader class to using the
SegmentReader. If
+ * null is passed, this behavior will be disabled and the stream reader
will be used.
+ *
+ * <p>Does not modify this object.
+ *
+ * <p>When we have a builder, we initialize the value. When they call the
method then we
+ * override the value
+ */
+ @Experimental(Kind.SOURCE_SINK)
+ public Read withMaxBufferElementCount(@Nullable Integer
maxBufferElementCount) {
Review Comment:
I think that would be a great idea, what file contains the integration tests
for BigtableIO? I should be able to get it done before the end of this half.
Issue Time Tracking
-------------------
Worklog Id: (was: 775567)
Time Spent: 2h (was: 1h 50m)
> Create separate BigtableIO reader that splits large scans into smaller
> segments
> -------------------------------------------------------------------------------
>
> Key: BEAM-14502
> URL: https://issues.apache.org/jira/browse/BEAM-14502
> Project: Beam
> Issue Type: Improvement
> Components: io-java-gcp
> Reporter: Diego E Gomez
> Assignee: Diego E Gomez
> Priority: P2
> Time Spent: 2h
> Remaining Estimate: 0h
>
> A separate BigtableIO reader needs to be created to split large scans into
> smaller batches. This will be to avoid OOMs on the server-side.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
