From 667a822fd1472802f7203ccfe1a8c228dc755d68 Mon Sep 17 00:00:00 2001 From: Jiao Mingye Date: Wed, 19 Nov 2025 12:39:40 +0800 Subject: [PATCH] [core] Fix parquet performance regression in reader init --- .../apache/paimon/format/parquet/ParquetReaderFactory.java | 4 +++- .../java/org/apache/paimon/format/parquet/ParquetUtil.java | 3 ++- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetReaderFactory.java b/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetReaderFactory.java index b27c81e332ca..5611bda959cf 100644 --- a/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetReaderFactory.java +++ b/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetReaderFactory.java @@ -36,6 +36,7 @@ import org.apache.paimon.utils.Preconditions; import org.apache.parquet.ParquetReadOptions; +import org.apache.parquet.conf.PlainParquetConfiguration; import org.apache.parquet.filter2.compat.FilterCompat; import org.apache.parquet.hadoop.ParquetFileReader; import org.apache.parquet.hadoop.ParquetInputFormat; @@ -101,7 +102,8 @@ public ParquetReaderFactory( public FileRecordReader createReader(FormatReaderFactory.Context context) throws IOException { ParquetReadOptions.Builder builder = - ParquetReadOptions.builder().withRange(0, context.fileSize()); + ParquetReadOptions.builder(new PlainParquetConfiguration()) + .withRange(0, context.fileSize()); setReadOptions(builder); ParquetFileReader reader = diff --git a/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetUtil.java b/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetUtil.java index 0ec4fa162a4a..26b6a7b4e9a2 100644 --- a/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetUtil.java +++ b/paimon-format/src/main/java/org/apache/paimon/format/parquet/ParquetUtil.java @@ -26,6 +26,7 @@ import org.apache.parquet.ParquetReadOptions; import org.apache.parquet.column.statistics.Statistics; +import org.apache.parquet.conf.PlainParquetConfiguration; import org.apache.parquet.hadoop.ParquetFileReader; import org.apache.parquet.hadoop.metadata.BlockMetaData; import org.apache.parquet.hadoop.metadata.ColumnChunkMetaData; @@ -83,7 +84,7 @@ public static ParquetFileReader getParquetReader(FileIO fileIO, Path path, long throws IOException { return new ParquetFileReader( ParquetInputFile.fromPath(fileIO, path, length), - ParquetReadOptions.builder().build(), + ParquetReadOptions.builder(new PlainParquetConfiguration()).build(), null); }