surekhasaharan commented on a change in pull request #6451: Fix inconsistent
segment size(#6448)
URL: https://github.com/apache/incubator-druid/pull/6451#discussion_r224832133
##########
File path:
sql/src/main/java/org/apache/druid/sql/calcite/schema/SystemSchema.java
##########
@@ -211,14 +210,64 @@ public TableType getJdbcTableType()
final Iterator<Entry<DataSegment, SegmentMetadataHolder>>
availableSegmentEntries = availableSegmentMetadata.entrySet()
.iterator();
+ // in memory map to store segment data from available segments
+ final Map<String, PartialSegmentData> partialSegmentDataMap = new
HashMap<>();
+ for (SegmentMetadataHolder holder : availableSegmentMetadata.values()) {
+ final PartialSegmentData data = new PartialSegmentData(
+ holder.isAvailable(),
+ holder.isRealtime(),
+ holder.getNumReplicas(),
+ holder.getNumRows()
+ );
+ partialSegmentDataMap.put(holder.getSegmentId(), data);
+ }
+
//get published segments from coordinator
final JsonParserIterator<DataSegment> metadataSegments =
getMetadataSegments(
druidLeaderClient,
jsonMapper,
responseHandler
);
- Set<String> availableSegmentIds = new HashSet<>();
+ final Set<String> segmentsAlreadySeen = new HashSet<>();
+
+ //auth check for published segments
+ final CloseableIterator<DataSegment> authorizedPublishedSegments =
getAuthorizedPublishedSegments(
+ metadataSegments,
+ root
+ );
+ final FluentIterable<Object[]> publishedSegments = FluentIterable
+ .from(() -> authorizedPublishedSegments)
+ .transform(val -> {
+ try {
+ if (!segmentsAlreadySeen.contains(val.getIdentifier())) {
+ segmentsAlreadySeen.add(val.getIdentifier());
+ }
+ final PartialSegmentData partialSegmentData =
partialSegmentDataMap.get(val.getIdentifier());
+ return new Object[]{
+ val.getIdentifier(),
+ val.getDataSource(),
+ val.getInterval().getStart(),
+ val.getInterval().getEnd(),
+ val.getSize(),
+ val.getVersion(),
+ val.getShardSpec().getPartitionNum(),
+ partialSegmentData == null ? 0L :
partialSegmentData.getNumReplicas(),
+ partialSegmentData == null ? -1L :
partialSegmentData.getNumRows(),
Review comment:
That's a good point, so `numRows` can be null, but since longs are not
nullable yet, so I was passing `-1L` (my thought was if I keep it 0, it may be
misleading that a segment has 0 rows, didn't think about the affect on sum),
but now I think, `0L` might make more sense if we treat missing values as zeros.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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]
