dnhatn commented on a change in pull request #1263: LUCENE-9228: Sort dv
updates by terms by applying
URL: https://github.com/apache/lucene-solr/pull/1263#discussion_r380734250
##########
File path:
lucene/core/src/java/org/apache/lucene/index/FrozenBufferedUpdates.java
##########
@@ -808,10 +808,17 @@ private void setField(String field) throws IOException {
} else {
termsEnum = null;
}
+ currentTermVisited = false;
}
}
- DocIdSetIterator nextTerm(String field, BytesRef term) throws IOException {
+ /**
+ * Returns a {@link DocIdSetIterator} for the given field and term if
exists. If the parameter
+ * {@code skipIfTermVisited} is true, then this method will return null if
the given and field
+ * was visited already. This optimization is currently used in
+ * {@link #applyDocValuesUpdates(BufferedUpdatesStream.SegmentState, Map,
long, boolean)}
+ */
+ DocIdSetIterator nextTerm(String field, BytesRef term, boolean
skipIfTermVisited) throws IOException {
Review comment:
It's doable and more contained than the current approach, but there are two
downsides with it. We need to compare the current term and the last term. Here,
we don't have to as we re-use the comparison between the reader term and the
current term. We also need to maintain an extra iterator for the last term as
we set BytesRef when iterating terms. I am happy to make this change if you
think the comparison cost is not significant.
----------------------------------------------------------------
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]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
