Avoid forcing conversion to String when writing RCBinary

The implementation of the RCBinary serializer makes an unnecessary call to StringObjectInspector.getPrimitiveJavaObject(), which forces a conversion of the utf-8 bytes to a Java String
prestodb · Dec 29, 2015 · 72917de · 72917de
1 parent 294b442
commit 72917de
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 1 deletion.
diff --git a/pom.xml b/pom.xml
@@ -272,7 +272,7 @@
             <dependency>
                 <groupId>com.facebook.presto.hive</groupId>
                 <artifactId>hive-apache</artifactId>
-                <version>0.15</version>
+                <version>0.16</version>
             </dependency>
 
             <dependency>

diff --git a/presto-hive/src/main/java/com/facebook/presto/hive/HivePageSink.java b/presto-hive/src/main/java/com/facebook/presto/hive/HivePageSink.java
@@ -43,6 +43,7 @@
 import org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat;
 import org.apache.hadoop.hive.serde2.SerDeException;
 import org.apache.hadoop.hive.serde2.Serializer;
+import org.apache.hadoop.hive.serde2.columnar.OptimizedLazyBinaryColumnarSerde;
 import org.apache.hadoop.hive.serde2.objectinspector.SettableStructObjectInspector;
 import org.apache.hadoop.hive.serde2.objectinspector.StructField;
 import org.apache.hadoop.io.compress.CompressionCodec;
@@ -530,6 +531,9 @@ public HiveRecordWriter(
 
             fieldCount = fileColumnNames.size();
 
+            if (serDe.equals(org.apache.hadoop.hive.serde2.columnar.LazyBinaryColumnarSerDe.class.getName())) {
+                serDe = OptimizedLazyBinaryColumnarSerde.class.getName();
+            }
             serializer = initializeSerializer(conf, schema, serDe);
             recordWriter = HiveWriteUtils.createRecordWriter(new Path(writePath, fileName), conf, schema, outputFormat);