Preserve utf8 from Bolt

There is no need to directly convert string values coming from Bolt to `java.lang.String` directly, we can preserve it as a `UTF8StringValue` until we actually need the value. Currently this will not give much in performance except for queries like `RETURN 'hello world'`, since most string operations currently requires an actual string instance. However going forward we can be smarter about preserving the `UTF8StringValue`.
neo4j · Oct 19, 2017 · 1b4ab9f · 1b4ab9f
1 parent 2b82af3
commit 1b4ab9f
Show file tree

Hide file tree

Showing 3 changed files with 23 additions and 6 deletions.
diff --git a/community/bolt/src/main/java/org/neo4j/bolt/v1/messaging/Neo4jPack.java b/community/bolt/src/main/java/org/neo4j/bolt/v1/messaging/Neo4jPack.java
@@ -32,11 +32,11 @@
 import org.neo4j.bolt.v1.packstream.PackType;
 import org.neo4j.bolt.v1.runtime.Neo4jError;
 import org.neo4j.collection.primitive.PrimitiveLongIntKeyValueArray;
-import org.neo4j.kernel.impl.util.BaseToObjectValueWriter;
 import org.neo4j.graphdb.Node;
 import org.neo4j.graphdb.Relationship;
 import org.neo4j.graphdb.spatial.Point;
 import org.neo4j.kernel.api.exceptions.Status;
+import org.neo4j.kernel.impl.util.BaseToObjectValueWriter;
 import org.neo4j.values.AnyValue;
 import org.neo4j.values.AnyValueWriter;
 import org.neo4j.values.storable.TextArray;
@@ -427,7 +427,7 @@ public AnyValue unpack() throws IOException
             case BYTES:
                 return byteArray( unpackBytes() );
             case STRING:
-                return Values.stringValue( unpackString() );
+                return Values.utf8Value( unpackUTF8() );
             case INTEGER:
                 return Values.longValue( unpackLong() );
             case FLOAT:

diff --git a/community/bolt/src/test/java/org/neo4j/bolt/v1/messaging/Neo4jPackTest.java b/community/bolt/src/test/java/org/neo4j/bolt/v1/messaging/Neo4jPackTest.java
@@ -36,6 +36,7 @@
 import org.neo4j.kernel.api.exceptions.Status;
 import org.neo4j.kernel.impl.util.ValueUtils;
 import org.neo4j.values.AnyValue;
+import org.neo4j.values.storable.StringValue;
 import org.neo4j.values.storable.TextArray;
 import org.neo4j.values.storable.TextValue;
 import org.neo4j.values.storable.Values;
@@ -47,6 +48,7 @@
 import static org.hamcrest.CoreMatchers.equalTo;
 import static org.hamcrest.CoreMatchers.instanceOf;
 import static org.hamcrest.MatcherAssert.assertThat;
+import static org.hamcrest.Matchers.is;
 import static org.neo4j.bolt.v1.messaging.example.Edges.ALICE_KNOWS_BOB;
 import static org.neo4j.bolt.v1.messaging.example.Nodes.ALICE;
 import static org.neo4j.bolt.v1.messaging.example.Paths.ALL_PATHS;
@@ -252,6 +254,7 @@ public void shouldPackUtf8() throws IOException
 
         // When
         AnyValue unpacked = unpacked( output.bytes() );
+        assertThat( unpacked, is( instanceOf( StringValue.UTF8StringValue.class ) ));
 
         // Then
         assertThat( unpacked, equalTo( textValue ) );

diff --git a/community/values/src/main/java/org/neo4j/values/storable/StringValue.java b/community/values/src/main/java/org/neo4j/values/storable/StringValue.java
@@ -20,6 +20,7 @@
 package org.neo4j.values.storable;
 
 import java.nio.charset.StandardCharsets;
+import java.util.Arrays;
 
 import static java.lang.String.format;
 
@@ -120,11 +121,11 @@ public int length()
      * Just as a normal StringValue but is backed by a byte array and does string
      * serialization lazily.
       *
-      * TODO in this implementation most operation will actually load the string
-      * such as hashCode, length, equals etc. These could be implemented using
-      * the byte array directly
+      * TODO in this implementation most operations will actually load the string
+      * such as hashCode, length. These could be implemented using
+      * the byte array directly in later optimizations
      */
-    static final class UTF8StringValue extends StringValue
+    public static final class UTF8StringValue extends StringValue
     {
         private volatile String value;
         private final byte[] bytes;
@@ -145,6 +146,19 @@ public <E extends Exception> void writeTo( ValueWriter<E> writer ) throws E
             writer.writeUTF8( bytes, offset, length );
         }
 
+        @Override
+        public boolean equals( Value value )
+        {
+            if ( value instanceof UTF8StringValue )
+            {
+                return Arrays.equals( bytes, ((UTF8StringValue) value).bytes );
+            }
+            else
+            {
+                return super.equals( value );
+            }
+        }
+
         @Override
         String value()
         {