Use sequential reader for nested stored fields

jimczi · Oct 1, 2024 · ee68cdf · ee68cdf
1 parent 1cfb6e8
commit ee68cdf
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 5 deletions.
diff --git a/benchmarks/src/main/java/org/elasticsearch/benchmark/search/SourceLoaderBenchmark.java b/benchmarks/src/main/java/org/elasticsearch/benchmark/search/SourceLoaderBenchmark.java
@@ -0,0 +1,2 @@
+package org.elasticsearch.benchmark.search;public class SourceLoaderBenchmark {
+}
diff --git a/server/src/main/java/org/elasticsearch/index/fieldvisitor/StoredFieldLoader.java b/server/src/main/java/org/elasticsearch/index/fieldvisitor/StoredFieldLoader.java
@@ -53,17 +53,24 @@ public static StoredFieldLoader fromSpec(StoredFieldsSpec spec) {
         return create(spec.requiresSource(), spec.requiredStoredFields());
     }
 
+    public static StoredFieldLoader create(boolean loadSource, Set<String> fields) {
+        return create(loadSource, fields, false);
+    }
+
     /**
      * Creates a new StoredFieldLoader
-     * @param loadSource should this loader load the _source field
-     * @param fields     a set of additional fields the loader should load
+     *
+     * @param loadSource           indicates whether this loader should load the {@code _source} field.
+     * @param fields               a set of additional fields that the loader should load.
+     * @param forceSequentialReader if {@code true}, forces the use of a sequential leaf reader;
+     *                              otherwise, uses the heuristic defined in {@link StoredFieldLoader#reader(LeafReaderContext, int[])}.
      */
-    public static StoredFieldLoader create(boolean loadSource, Set<String> fields) {
+    public static StoredFieldLoader create(boolean loadSource, Set<String> fields, boolean forceSequentialReader) {
         List<String> fieldsToLoad = fieldsToLoad(loadSource, fields);
         return new StoredFieldLoader() {
             @Override
             public LeafStoredFieldLoader getLoader(LeafReaderContext ctx, int[] docs) throws IOException {
-                return new ReaderStoredFieldLoader(reader(ctx, docs), loadSource, fields);
+                return new ReaderStoredFieldLoader(forceSequentialReader ? sequentialReader(ctx) : reader(ctx, docs), loadSource, fields);
             }
 
             @Override

diff --git a/server/src/main/java/org/elasticsearch/index/mapper/NestedObjectMapper.java b/server/src/main/java/org/elasticsearch/index/mapper/NestedObjectMapper.java
@@ -403,7 +403,8 @@ SourceLoader.SyntheticFieldLoader syntheticFieldLoader(SourceFilter filter, Coll
         SourceLoader sourceLoader = new SourceLoader.Synthetic(() -> super.syntheticFieldLoader(filter, mappers, true), NOOP);
         // Some synthetic source use cases require using _ignored_source field
         var requiredStoredFields = IgnoredSourceFieldMapper.ensureLoaded(sourceLoader.requiredStoredFields(), indexSettings);
-        var storedFieldLoader = org.elasticsearch.index.fieldvisitor.StoredFieldLoader.create(false, requiredStoredFields);
+        // force sequential access since nested fields are indexed per block
+        var storedFieldLoader = org.elasticsearch.index.fieldvisitor.StoredFieldLoader.create(false, requiredStoredFields, true);
         return new NestedSyntheticFieldLoader(
             storedFieldLoader,
             sourceLoader,
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,2 @@
		package org.elasticsearch.benchmark.search;public class SourceLoaderBenchmark {
		}