[GitHub] leventov commented on a change in pull request #5913: Move Caching Cluster Client to java streams and allow parallel intermediate merges

[GitBox <gi...@apache.org>]

leventov commented on a change in pull request #5913: Move Caching Cluster Client to java streams and allow parallel intermediate merges
URL: https://github.com/apache/incubator-druid/pull/5913#discussion_r205995384
 
 

 ##########
 File path: server/src/main/java/io/druid/client/CachingClusteredClient.java
 ##########
 @@ -249,74 +296,88 @@ public CachingClusteredClient(
         contextBuilder.put(CacheConfig.POPULATE_CACHE, false);
         contextBuilder.put("bySegment", true);
       }
-      return contextBuilder.build();
+      return Collections.unmodifiableMap(contextBuilder);
     }
 
-    Sequence<T> run(final UnaryOperator<TimelineLookup<String, ServerSelector>> timelineConverter)
+    Stream<Sequence<T>> run(final UnaryOperator<TimelineLookup<String, ServerSelector>> timelineConverter)
     {
       @Nullable
       TimelineLookup<String, ServerSelector> timeline = serverView.getTimeline(query.getDataSource());
       if (timeline == null) {
-        return Sequences.empty();
+        return Stream.empty();
       }
       timeline = timelineConverter.apply(timeline);
       if (uncoveredIntervalsLimit > 0) {
         computeUncoveredIntervals(timeline);
       }
 
-      final Set<ServerToSegment> segments = computeSegmentsToQuery(timeline);
+      Stream<ServerToSegment> segments = computeSegmentsToQuery(timeline);
       @Nullable
       final byte[] queryCacheKey = computeQueryCacheKey();
       if (query.getContext().get(QueryResource.HEADER_IF_NONE_MATCH) != null) {
+        // Materialize then re-stream
+        List<ServerToSegment> materializedSegments = segments.collect(Collectors.toList());
+        segments = materializedSegments.stream();
+
         @Nullable
         final String prevEtag = (String) query.getContext().get(QueryResource.HEADER_IF_NONE_MATCH);
         @Nullable
-        final String currentEtag = computeCurrentEtag(segments, queryCacheKey);
+        final String currentEtag = computeCurrentEtag(materializedSegments, queryCacheKey);
         if (currentEtag != null && currentEtag.equals(prevEtag)) {
-          return Sequences.empty();
+          return Stream.empty();
         }
       }
 
-      final List<Pair<Interval, byte[]>> alreadyCachedResults = pruneSegmentsWithCachedResults(queryCacheKey, segments);
-      final SortedMap<DruidServer, List<SegmentDescriptor>> segmentsByServer = groupSegmentsByServer(segments);
-      return new LazySequence<>(() -> {
-        List<Sequence<T>> sequencesByInterval = new ArrayList<>(alreadyCachedResults.size() + segmentsByServer.size());
-        addSequencesFromCache(sequencesByInterval, alreadyCachedResults);
-        addSequencesFromServer(sequencesByInterval, segmentsByServer);
-        return Sequences
-            .simple(sequencesByInterval)
-            .flatMerge(seq -> seq, query.getResultOrdering());
-      });
+      // This pipeline follows a few general steps:
+      // 1. Fetch cache results - Unfortunately this is an eager operation so that the non cached items can
+      // be batched per server. Cached results are assigned to a mock server ALREADY_CACHED_SERVER
+      // 2. Group the segment information by server
+      // 3. Per server (including the ALREADY_CACHED_SERVER) create the appropriate Sequence results - cached results
+      // are handled in their own merge
+      final Stream<SerializablePair<ServerToSegment, Optional<T>>> cacheResolvedResults = deserializeFromCache(
+          maybeFetchCacheResults(
+              queryCacheKey,
+              segments
+          )
+      );
+      return groupCachedResultsByServer(
 
 Review comment:
   The better formatting for you snippet is
   ```java
     private Stream<ServerToSegment> computeSegmentsToQuery(TimelineLookup<String, ServerSelector> timeline)
     {
       return toolChest
           .filterSegments(
               query,
               query.getIntervals().stream().flatMap(i -> timeline.lookup(i).stream()).collect(Collectors.toList())
           )
           .stream()
           .flatMap(holder -> DimFilterUtils
               .filterShards(
                   query.getFilter(),
                   holder.getObject(),
                   partitionChunk -> partitionChunk.getObject().getSegment().getShardSpec(),
                   Maps.newHashMap()
               )
               .stream()
               .map(chunk -> {
                 ServerSelector server = chunk.getObject();
                 final SegmentDescriptor segment = new SegmentDescriptor(
                     holder.getInterval(),
                     holder.getVersion(),
                     chunk.getChunkNumber()
                 );
                 return new ServerToSegment(server, segment);
               })
           )
           .distinct();
     }
   ```
   
   However, note that "Effective Java" recommends to extract every lambda longer than 3 lines as a method. The above snippet contains two lambdas much longer than that: one of 16 lines and another of 7 lines. If they are extracted as methods,  Stream API code will look very accurate.
   
   Another point: new IntelliJ [added type hints for Stream code](https://twitter.com/antonarhipov/status/1022139362736304129), that won't work in your formatting style, they only work in the "classical" formatting style.
   
   Also, the classical style is predominately used in the Druid codebase already, e. g. in Query building code in tests.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org