Optimization in fielddata cache where ordinals are used instead of flat arrays when number of unique values is low

src/main/java/org/elasticsearch/index/fielddata/plain/DoubleArrayIndexFieldData.java

@@ -29,6 +29,7 @@ import org.apache.lucene.util.FixedBitSet;
 import org.apache.lucene.util.NumericUtils;
 import org.elasticsearch.ElasticSearchException;
 import org.elasticsearch.common.Nullable;
+import org.elasticsearch.common.RamUsage;
 import org.elasticsearch.common.settings.Settings;
 import org.elasticsearch.index.Index;
 import org.elasticsearch.index.fielddata.*;
@@ -103,12 +104,22 @@ public class DoubleArrayIndexFieldData extends AbstractIndexFieldData<DoubleArra
             Ordinals build = builder.build(fieldDataType.getSettings());
             if (!build.isMultiValued() && CommonSettings.removeOrdsOnSingleValue(fieldDataType)) {
                 Docs ordinals = build.ordinals();
+                final FixedBitSet set = builder.buildDocsWithValuesSet();
+
+                // there's sweatspot where due to low unique value count, using ordinals will consume less memory
+                long singleValuesArraySize = reader.maxDoc() * RamUsage.NUM_BYTES_DOUBLE + (set == null ? 0 : set.getBits().length * RamUsage.NUM_BYTES_LONG + RamUsage.NUM_BYTES_INT);
+                long uniqueValuesArraySize = values.size() * RamUsage.NUM_BYTES_DOUBLE;
+                long ordinalsSize = build.getMemorySizeInBytes();
+                if (uniqueValuesArraySize + ordinalsSize < singleValuesArraySize) {
+                    return new DoubleArrayAtomicFieldData.WithOrdinals(values.toArray(new double[values.size()]), reader.maxDoc(), build);
+                }
+
                 double[] sValues = new double[reader.maxDoc()];
                 int maxDoc = reader.maxDoc();
                 for (int i = 0; i < maxDoc; i++) {
                     sValues[i] = values.get(ordinals.getOrd(i));
                 }
-                final FixedBitSet set = builder.buildDocsWithValuesSet();
+
                 if (set == null) {
                     return new DoubleArrayAtomicFieldData.Single(sValues, reader.maxDoc());
                 } else {

src/main/java/org/elasticsearch/index/fielddata/plain/FloatArrayIndexFieldData.java

@@ -29,6 +29,7 @@ import org.apache.lucene.util.FixedBitSet;
 import org.apache.lucene.util.NumericUtils;
 import org.elasticsearch.ElasticSearchException;
 import org.elasticsearch.common.Nullable;
+import org.elasticsearch.common.RamUsage;
 import org.elasticsearch.common.settings.Settings;
 import org.elasticsearch.index.Index;
 import org.elasticsearch.index.fielddata.*;
@@ -103,12 +104,21 @@ public class FloatArrayIndexFieldData extends AbstractIndexFieldData<FloatArrayA
             Ordinals build = builder.build(fieldDataType.getSettings());
             if (!build.isMultiValued() && CommonSettings.removeOrdsOnSingleValue(fieldDataType)) {
                 Docs ordinals = build.ordinals();
+                final FixedBitSet set = builder.buildDocsWithValuesSet();
+
+                // there's sweatspot where due to low unique value count, using ordinals will consume less memory
+                long singleValuesArraySize = reader.maxDoc() * RamUsage.NUM_BYTES_FLOAT + (set == null ? 0 : set.getBits().length * RamUsage.NUM_BYTES_LONG + RamUsage.NUM_BYTES_INT);
+                long uniqueValuesArraySize = values.size() * RamUsage.NUM_BYTES_FLOAT;
+                long ordinalsSize = build.getMemorySizeInBytes();
+                if (uniqueValuesArraySize + ordinalsSize < singleValuesArraySize) {
+                    return new FloatArrayAtomicFieldData.WithOrdinals(values.toArray(new float[values.size()]), reader.maxDoc(), build);
+                }
+
                 float[] sValues = new float[reader.maxDoc()];
                 int maxDoc = reader.maxDoc();
                 for (int i = 0; i < maxDoc; i++) {
                     sValues[i] = values.get(ordinals.getOrd(i));
                 }
-                final FixedBitSet set = builder.buildDocsWithValuesSet();
                 if (set == null) {
                     return new FloatArrayAtomicFieldData.Single(sValues, reader.maxDoc());
                 } else {

src/main/java/org/elasticsearch/index/fielddata/plain/IntArrayIndexFieldData.java

@@ -30,6 +30,7 @@ import org.apache.lucene.util.FixedBitSet;
 import org.apache.lucene.util.NumericUtils;
 import org.elasticsearch.ElasticSearchException;
 import org.elasticsearch.common.Nullable;
+import org.elasticsearch.common.RamUsage;
 import org.elasticsearch.common.settings.Settings;
 import org.elasticsearch.index.Index;
 import org.elasticsearch.index.fielddata.*;
@@ -146,6 +147,11 @@ public class IntArrayIndexFieldData extends AbstractIndexFieldData<AtomicNumeric
                             }
                             return sValues;
                         }
+
+                        @Override
+                        public int size() {
+                            return values.size();
+                        }
                     });
                 }
             }
@@ -160,6 +166,11 @@ public class IntArrayIndexFieldData extends AbstractIndexFieldData<AtomicNumeric
                 public int[] toArray() {
                     return values.toArray();
                 }
+
+                @Override
+                public int size() {
+                    return values.size();
+                }
             });
         } finally {
             builder.close();
@@ -170,17 +181,28 @@ public class IntArrayIndexFieldData extends AbstractIndexFieldData<AtomicNumeric
         int get(int index);
 
         int[] toArray();
+
+        int size();
     }
 
     static IntArrayAtomicFieldData build(AtomicReader reader, FieldDataType fieldDataType, OrdinalsBuilder builder, Ordinals build, BuilderIntegers values) {
         if (!build.isMultiValued() && CommonSettings.removeOrdsOnSingleValue(fieldDataType)) {
             Docs ordinals = build.ordinals();
+            final FixedBitSet set = builder.buildDocsWithValuesSet();
+
+            // there's sweatspot where due to low unique value count, using ordinals will consume less memory
+            long singleValuesArraySize = reader.maxDoc() * RamUsage.NUM_BYTES_INT + (set == null ? 0 : set.getBits().length * RamUsage.NUM_BYTES_LONG + RamUsage.NUM_BYTES_INT);
+            long uniqueValuesArraySize = values.size() * RamUsage.NUM_BYTES_INT;
+            long ordinalsSize = build.getMemorySizeInBytes();
+            if (uniqueValuesArraySize + ordinalsSize < singleValuesArraySize) {
+                return new IntArrayAtomicFieldData.WithOrdinals(values.toArray(), reader.maxDoc(), build);
+            }
+
             int[] sValues = new int[reader.maxDoc()];
             int maxDoc = reader.maxDoc();
             for (int i = 0; i < maxDoc; i++) {
                 sValues[i] = values.get(ordinals.getOrd(i));
             }
-            final FixedBitSet set = builder.buildDocsWithValuesSet();
             if (set == null) {
                 return new IntArrayAtomicFieldData.Single(sValues, reader.maxDoc());
             } else {

src/main/java/org/elasticsearch/index/fielddata/plain/LongArrayIndexFieldData.java

@@ -30,6 +30,7 @@ import org.apache.lucene.util.FixedBitSet;
 import org.apache.lucene.util.NumericUtils;
 import org.elasticsearch.ElasticSearchException;
 import org.elasticsearch.common.Nullable;
+import org.elasticsearch.common.RamUsage;
 import org.elasticsearch.common.settings.Settings;
 import org.elasticsearch.index.Index;
 import org.elasticsearch.index.fielddata.*;
@@ -146,6 +147,11 @@ public class LongArrayIndexFieldData extends AbstractIndexFieldData<AtomicNumeri
                             }
                             return sValues;
                         }
+
+                        @Override
+                        public int size() {
+                            return values.size();
+                        }
                     });
                 } else if (min >= Integer.MIN_VALUE && max <= Integer.MAX_VALUE) {
                     return IntArrayIndexFieldData.build(reader, fieldDataType, builder, build, new IntArrayIndexFieldData.BuilderIntegers() {
@@ -163,18 +169,33 @@ public class LongArrayIndexFieldData extends AbstractIndexFieldData<AtomicNumeri
                             }
                             return iValues;
                         }
+
+                        @Override
+                        public int size() {
+                            return values.size();
+                        }
+
                     });
                 }
             }
 
             if (!build.isMultiValued() && CommonSettings.removeOrdsOnSingleValue(fieldDataType)) {
                 Docs ordinals = build.ordinals();
+                final FixedBitSet set = builder.buildDocsWithValuesSet();
+
+                // there's sweatspot where due to low unique value count, using ordinals will consume less memory
+                long singleValuesArraySize = reader.maxDoc() * RamUsage.NUM_BYTES_LONG + (set == null ? 0 : set.getBits().length * RamUsage.NUM_BYTES_LONG + RamUsage.NUM_BYTES_INT);
+                long uniqueValuesArraySize = values.size() * RamUsage.NUM_BYTES_LONG;
+                long ordinalsSize = build.getMemorySizeInBytes();
+                if (uniqueValuesArraySize + ordinalsSize < singleValuesArraySize) {
+                    return new LongArrayAtomicFieldData.WithOrdinals(values.toArray(new long[values.size()]), reader.maxDoc(), build);
+                }
+
                 long[] sValues = new long[reader.maxDoc()];
                 int maxDoc = reader.maxDoc();
                 for (int i = 0; i < maxDoc; i++) {
                     sValues[i] = values.get(ordinals.getOrd(i));
                 }
-                final FixedBitSet set = builder.buildDocsWithValuesSet();
                 if (set == null) {
                     return new LongArrayAtomicFieldData.Single(sValues, reader.maxDoc());
                 } else {

src/main/java/org/elasticsearch/index/fielddata/plain/ShortArrayIndexFieldData.java

@@ -30,6 +30,7 @@ import org.apache.lucene.util.FixedBitSet;
 import org.apache.lucene.util.NumericUtils;
 import org.elasticsearch.ElasticSearchException;
 import org.elasticsearch.common.Nullable;
+import org.elasticsearch.common.RamUsage;
 import org.elasticsearch.common.settings.Settings;
 import org.elasticsearch.index.Index;
 import org.elasticsearch.index.fielddata.*;
@@ -143,6 +144,11 @@ public class ShortArrayIndexFieldData extends AbstractIndexFieldData<AtomicNumer
                 public short[] toArray() {
                     return values.toArray();
                 }
+
+                @Override
+                public int size() {
+                    return values.size();
+                }
             });
         } finally {
             builder.close();
@@ -153,17 +159,29 @@ public class ShortArrayIndexFieldData extends AbstractIndexFieldData<AtomicNumer
         short get(int index);
 
         short[] toArray();
+
+        int size();
     }
 
     static ShortArrayAtomicFieldData build(AtomicReader reader, FieldDataType fieldDataType, OrdinalsBuilder builder, Ordinals build, BuilderShorts values) {
         if (!build.isMultiValued() && CommonSettings.removeOrdsOnSingleValue(fieldDataType)) {
             Docs ordinals = build.ordinals();
+            final FixedBitSet set = builder.buildDocsWithValuesSet();
+
+            // there's sweatspot where due to low unique value count, using ordinals will consume less memory
+            long singleValuesArraySize = reader.maxDoc() * RamUsage.NUM_BYTES_SHORT + (set == null ? 0 : set.getBits().length * RamUsage.NUM_BYTES_LONG + RamUsage.NUM_BYTES_INT);
+            long uniqueValuesArraySize = values.size() * RamUsage.NUM_BYTES_SHORT;
+            long ordinalsSize = build.getMemorySizeInBytes();
+            if (uniqueValuesArraySize + ordinalsSize < singleValuesArraySize) {
+                return new ShortArrayAtomicFieldData.WithOrdinals(values.toArray(), reader.maxDoc(), build);
+            }
+
             short[] sValues = new short[reader.maxDoc()];
             int maxDoc = reader.maxDoc();
             for (int i = 0; i < maxDoc; i++) {
                 sValues[i] = values.get(ordinals.getOrd(i));
             }
-            final FixedBitSet set = builder.buildDocsWithValuesSet();
+
             if (set == null) {
                 return new ShortArrayAtomicFieldData.Single(sValues, reader.maxDoc());
             } else {