(paimon) branch master updated: [parquet] Optimize parquet to use getBytesUnsafe for dictionary (#3468)

[junhao]

This is an automated email from the ASF dual-hosted git repository.

junhao pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/paimon.git


The following commit(s) were added to refs/heads/master by this push:
     new 88aa434aa [parquet] Optimize parquet to use getBytesUnsafe for 
dictionary (#3468)
88aa434aa is described below

commit 88aa434aad2321226d6a1c927bd3db687d3e5262
Author: Jingsong Lee <jingsongl...@gmail.com>
AuthorDate: Tue Jun 4 15:16:58 2024 +0800

    [parquet] Optimize parquet to use getBytesUnsafe for dictionary (#3468)
---
 .../java/org/apache/paimon/format/parquet/reader/BytesColumnReader.java | 2 +-
 .../apache/paimon/format/parquet/reader/FixedLenBytesColumnReader.java  | 2 +-
 .../java/org/apache/paimon/format/parquet/reader/ParquetDictionary.java | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)

diff --git 
a/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/BytesColumnReader.java
 
b/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/BytesColumnReader.java
index da64d7db5..e83115c8a 100644
--- 
a/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/BytesColumnReader.java
+++ 
b/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/BytesColumnReader.java
@@ -75,7 +75,7 @@ public class BytesColumnReader extends 
AbstractColumnReader<WritableBytesVector>
             int rowId, int num, WritableBytesVector column, WritableIntVector 
dictionaryIds) {
         for (int i = rowId; i < rowId + num; ++i) {
             if (!column.isNullAt(i)) {
-                byte[] bytes = 
dictionary.decodeToBinary(dictionaryIds.getInt(i)).getBytes();
+                byte[] bytes = 
dictionary.decodeToBinary(dictionaryIds.getInt(i)).getBytesUnsafe();
                 column.appendBytes(i, bytes, 0, bytes.length);
             }
         }
diff --git 
a/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/FixedLenBytesColumnReader.java
 
b/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/FixedLenBytesColumnReader.java
index de001065e..1fc9d9e91 100644
--- 
a/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/FixedLenBytesColumnReader.java
+++ 
b/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/FixedLenBytesColumnReader.java
@@ -70,7 +70,7 @@ public class FixedLenBytesColumnReader<VECTOR extends 
WritableColumnVector>
             WritableBytesVector bytesVector = (WritableBytesVector) column;
             for (int i = 0; i < num; i++) {
                 if (runLenDecoder.readInteger() == maxDefLevel) {
-                    byte[] bytes = readDataBinary(bytesLen).getBytes();
+                    byte[] bytes = readDataBinary(bytesLen).getBytesUnsafe();
                     bytesVector.appendBytes(rowId + i, bytes, 0, bytes.length);
                 } else {
                     bytesVector.setNullAt(rowId + i);
diff --git 
a/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/ParquetDictionary.java
 
b/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/ParquetDictionary.java
index cc636f080..a8deafa0f 100644
--- 
a/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/ParquetDictionary.java
+++ 
b/paimon-format/src/main/java/org/apache/paimon/format/parquet/reader/ParquetDictionary.java
@@ -54,7 +54,7 @@ public final class ParquetDictionary implements Dictionary {
 
     @Override
     public byte[] decodeToBinary(int id) {
-        return dictionary.decodeToBinary(id).getBytes();
+        return dictionary.decodeToBinary(id).getBytesUnsafe();
     }
 
     @Override