Fix match_only_text for keyword multi-fields with ignore_above (#131314) (#131338)

(cherry picked from commit 6f8be9c)

# Conflicts:
#	server/src/main/java/org/elasticsearch/index/mapper/KeywordFieldMapper.java

Author: jordan-powers

modules/mapper-extras/src/main/java/org/elasticsearch/index/mapper/extras/MatchOnlyTextFieldMapper.java

@@ -47,6 +47,7 @@
 import org.elasticsearch.index.mapper.BlockStoredFieldsReader;
 import org.elasticsearch.index.mapper.DocumentParserContext;
 import org.elasticsearch.index.mapper.FieldMapper;
+import org.elasticsearch.index.mapper.KeywordFieldMapper;
 import org.elasticsearch.index.mapper.MappedFieldType;
 import org.elasticsearch.index.mapper.MapperBuilderContext;
 import org.elasticsearch.index.mapper.SourceValueFetcher;
@@ -66,6 +67,7 @@
 import java.io.IOException;
 import java.io.UncheckedIOException;
 import java.util.ArrayList;
+import java.util.Arrays;
 import java.util.Collections;
 import java.util.List;
 import java.util.Map;
@@ -246,6 +248,10 @@ private IOFunction<LeafReaderContext, CheckedIntFunction<List<Object>, IOExcepti
                 String parentField = searchExecutionContext.parentPath(name());
                 var parent = searchExecutionContext.lookup().fieldType(parentField);
                 if (parent.isStored()) {
+                    if (parent instanceof KeywordFieldMapper.KeywordFieldType keywordParent
+                        && keywordParent.ignoreAbove() != Integer.MAX_VALUE) {
+                        return storedFieldFetcher(parentField, keywordParent.originalName());
+                    }
                     return storedFieldFetcher(parentField);
                 } else if (parent.hasDocValues()) {
                     var ifd = searchExecutionContext.getForField(parent, MappedFieldType.FielddataOperation.SEARCH);
@@ -259,7 +265,11 @@ private IOFunction<LeafReaderContext, CheckedIntFunction<List<Object>, IOExcepti
                 if (kwd != null) {
                     var fieldType = kwd.fieldType();
                     if (fieldType.isStored()) {
-                        return storedFieldFetcher(fieldType.name());
+                        if (fieldType.ignoreAbove() != Integer.MAX_VALUE) {
+                            return storedFieldFetcher(fieldType.name(), fieldType.originalName());
+                        } else {
+                            return storedFieldFetcher(fieldType.name());
+                        }
                     } else if (fieldType.hasDocValues()) {
                         var ifd = searchExecutionContext.getForField(fieldType, MappedFieldType.FielddataOperation.SEARCH);
                         return docValuesFieldFetcher(ifd);
@@ -306,13 +316,17 @@ private static IOFunction<LeafReaderContext, CheckedIntFunction<List<Object>, IO
             };
         }
 
-        private static IOFunction<LeafReaderContext, CheckedIntFunction<List<Object>, IOException>> storedFieldFetcher(String name) {
-            var loader = StoredFieldLoader.create(false, Set.of(name));
+        private static IOFunction<LeafReaderContext, CheckedIntFunction<List<Object>, IOException>> storedFieldFetcher(String... names) {
+            var loader = StoredFieldLoader.create(false, Set.of(names));
             return context -> {
                 var leafLoader = loader.getLoader(context, null);
                 return docId -> {
                     leafLoader.advanceTo(docId);
-                    return leafLoader.storedFields().get(name);
+                    var storedFields = leafLoader.storedFields();
+                    if (names.length == 1) {
+                        return storedFields.get(names[0]);
+                    }
+                    return Arrays.stream(names).map(storedFields::get).filter(Objects::nonNull).flatMap(List::stream).toList();
                 };
             };
         }

modules/mapper-extras/src/yamlRestTest/resources/rest-api-spec/test/match_only_text/10_basic.yml

@@ -479,6 +479,49 @@ synthetic_source match_only_text as multi-field with stored keyword as parent:
       hits.hits.0._source.foo: "Apache Lucene powers Elasticsearch"
 
 ---
+synthetic_source match_only_text as multi-field with ignored stored keyword as parent:
+  - requires:
+      cluster_features: [ "mapper.source.mode_from_index_setting" ]
+      reason: "Source mode configured through index setting"
+
+  - do:
+      indices.create:
+        index: synthetic_source_test
+        body:
+          settings:
+            index:
+              mapping.source.mode: synthetic
+          mappings:
+            properties:
+              foo:
+                type: keyword
+                store: true
+                doc_values: false
+                ignore_above: 10
+                fields:
+                  text:
+                    type: match_only_text
+
+  - do:
+      index:
+        index: synthetic_source_test
+        id: "1"
+        refresh: true
+        body:
+          foo: "Apache Lucene powers Elasticsearch"
+
+  - do:
+      search:
+        index: synthetic_source_test
+        body:
+          query:
+            match_phrase:
+              foo.text: apache lucene
+
+  - match: { "hits.total.value": 1 }
+  - match:
+      hits.hits.0._source.foo: "Apache Lucene powers Elasticsearch"
+---
 synthetic_source match_only_text with multi-field:
   - requires:
       cluster_features: [ "mapper.source.mode_from_index_setting" ]
@@ -561,3 +604,47 @@ synthetic_source match_only_text with stored multi-field:
   - match: { "hits.total.value": 1 }
   - match:
       hits.hits.0._source.foo: "Apache Lucene powers Elasticsearch"
+
+---
+synthetic_source match_only_text with ignored stored multi-field:
+  - requires:
+      cluster_features: [ "mapper.source.mode_from_index_setting" ]
+      reason: "Source mode configured through index setting"
+
+  - do:
+      indices.create:
+        index: synthetic_source_test
+        body:
+          settings:
+            index:
+              mapping.source.mode: synthetic
+          mappings:
+            properties:
+              foo:
+                type: match_only_text
+                fields:
+                  raw:
+                    type: keyword
+                    store: true
+                    doc_values: false
+                    ignore_above: 10
+
+  - do:
+      index:
+        index: synthetic_source_test
+        id: "1"
+        refresh: true
+        body:
+          foo: "Apache Lucene powers Elasticsearch"
+
+  - do:
+      search:
+        index: synthetic_source_test
+        body:
+          query:
+            match_phrase:
+              foo: apache lucene
+
+  - match: { "hits.total.value": 1 }
+  - match:
+      hits.hits.0._source.foo: "Apache Lucene powers Elasticsearch"

server/src/main/java/org/elasticsearch/index/mapper/KeywordFieldMapper.java

@@ -425,6 +425,7 @@ public static final class KeywordFieldType extends StringFieldType {
         private final FieldValues<String> scriptValues;
         private final boolean isDimension;
         private final boolean isSyntheticSource;
+        private final String originalName;
 
         public KeywordFieldType(
             String name,
@@ -450,6 +451,7 @@ public KeywordFieldType(
             this.scriptValues = builder.scriptValues();
             this.isDimension = builder.dimension.getValue();
             this.isSyntheticSource = isSyntheticSource;
+            this.originalName = isSyntheticSource ? name + "._original" : null;
         }
 
         public KeywordFieldType(String name, boolean isIndexed, boolean hasDocValues, Map<String, String> meta) {
@@ -461,6 +463,7 @@ public KeywordFieldType(String name, boolean isIndexed, boolean hasDocValues, Ma
             this.scriptValues = null;
             this.isDimension = false;
             this.isSyntheticSource = false;
+            this.originalName = null;
         }
 
         public KeywordFieldType(String name) {
@@ -483,6 +486,7 @@ public KeywordFieldType(String name, FieldType fieldType) {
             this.scriptValues = null;
             this.isDimension = false;
             this.isSyntheticSource = false;
+            this.originalName = null;
         }
 
         public KeywordFieldType(String name, NamedAnalyzer analyzer) {
@@ -494,6 +498,7 @@ public KeywordFieldType(String name, NamedAnalyzer analyzer) {
             this.scriptValues = null;
             this.isDimension = false;
             this.isSyntheticSource = false;
+            this.originalName = null;
         }
 
         @Override
@@ -945,6 +950,15 @@ public Query automatonQuery(
         ) {
             return new AutomatonQueryWithDescription(new Term(name()), automatonSupplier.get(), description);
         }
+
+        /**
+         * The name used to store "original" that have been ignored
+         * by {@link KeywordFieldType#ignoreAbove()} so that they can be rebuilt
+         * for synthetic source.
+         */
+        public String originalName() {
+            return originalName;
+        }
     }
 
     private final boolean indexed;
@@ -992,7 +1006,7 @@ private KeywordFieldMapper(
         this.ignoreAbove = builder.ignoreAbove.getValue();
         this.offsetsFieldName = offsetsFieldName;
         this.indexSourceKeepMode = indexSourceKeepMode;
-        this.originalName = isSyntheticSource ? fullPath() + "._original" : null;
+        this.originalName = mappedFieldType.originalName();
     }
 
     @Override
@@ -1052,7 +1066,7 @@ private boolean indexValue(DocumentParserContext context, XContentString value)
                 // Save a copy of the field so synthetic source can load it
                 var utfBytes = value.bytes();
                 var bytesRef = new BytesRef(utfBytes.bytes(), utfBytes.offset(), utfBytes.length());
-                context.doc().add(new StoredField(originalName(), bytesRef));
+                context.doc().add(new StoredField(originalName, bytesRef));
             }
             return false;
         }
@@ -1155,15 +1169,6 @@ boolean hasNormalizer() {
         return normalizerName != null;
     }
 
-    /**
-     * The name used to store "original" that have been ignored
-     * by {@link KeywordFieldType#ignoreAbove()} so that they can be rebuilt
-     * for synthetic source.
-     */
-    private String originalName() {
-        return originalName;
-    }
-
     @Override
     protected SyntheticSourceSupport syntheticSourceSupport() {
         if (hasNormalizer()) {
@@ -1212,7 +1217,7 @@ protected BytesRef preserve(BytesRef value) {
         }
 
         if (fieldType().ignoreAbove != Integer.MAX_VALUE) {
-            layers.add(new CompositeSyntheticFieldLoader.StoredFieldLayer(originalName()) {
+            layers.add(new CompositeSyntheticFieldLoader.StoredFieldLayer(originalName) {
                 @Override
                 protected void writeValue(Object value, XContentBuilder b) throws IOException {
                     BytesRef ref = (BytesRef) value;