From 851e247caad0977cfd4998254d9602624e06539f Mon Sep 17 00:00:00 2001
From: Weizhong Lin <linweizhong@huawei.com>
Date: Wed, 8 Jul 2015 22:18:39 -0700
Subject: [PATCH] [SPARK-8928] [SQL] Makes CatalystSchemaConverter sticking to
1.4.x- when handling Parquet LISTs in compatible mode
This PR is based on #7209 authored by Sephiroth-Lin.
Author: Weizhong Lin <linweizhong@huawei.com>
Closes #7314 from liancheng/spark-8928 and squashes the following commits:
75267fe [Cheng Lian] Makes CatalystSchemaConverter sticking to 1.4.x- when handling LISTs in compatible mode
---
.../spark/sql/parquet/CatalystSchemaConverter.scala | 6 ++++--
.../apache/spark/sql/parquet/ParquetSchemaSuite.scala | 10 +++++-----
2 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/sql/core/src/main/scala/org/apache/spark/sql/parquet/CatalystSchemaConverter.scala b/sql/core/src/main/scala/org/apache/spark/sql/parquet/CatalystSchemaConverter.scala
index de3a72d814..1ea6926af6 100644
--- a/sql/core/src/main/scala/org/apache/spark/sql/parquet/CatalystSchemaConverter.scala
+++ b/sql/core/src/main/scala/org/apache/spark/sql/parquet/CatalystSchemaConverter.scala
@@ -461,7 +461,8 @@ private[parquet] class CatalystSchemaConverter(
field.name,
Types
.buildGroup(REPEATED)
- .addField(convertField(StructField("element", elementType, nullable)))
+ // "array_element" is the name chosen by parquet-hive (1.7.0 and prior version)
+ .addField(convertField(StructField("array_element", elementType, nullable)))
.named(CatalystConverter.ARRAY_CONTAINS_NULL_BAG_SCHEMA_NAME))
// Spark 1.4.x and prior versions convert ArrayType with non-nullable elements into a 2-level
@@ -474,7 +475,8 @@ private[parquet] class CatalystSchemaConverter(
ConversionPatterns.listType(
repetition,
field.name,
- convertField(StructField("element", elementType, nullable), REPEATED))
+ // "array" is the name chosen by parquet-avro (1.7.0 and prior version)
+ convertField(StructField("array", elementType, nullable), REPEATED))
// Spark 1.4.x and prior versions convert MapType into a 3-level group annotated by
// MAP_KEY_VALUE. This is covered by `convertGroupField(field: GroupType): DataType`.
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/parquet/ParquetSchemaSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/parquet/ParquetSchemaSuite.scala
index 35d3c33f99..fa62939267 100644
--- a/sql/core/src/test/scala/org/apache/spark/sql/parquet/ParquetSchemaSuite.scala
+++ b/sql/core/src/test/scala/org/apache/spark/sql/parquet/ParquetSchemaSuite.scala
@@ -174,7 +174,7 @@ class ParquetSchemaInferenceSuite extends ParquetSchemaTest {
"""
|message root {
| optional group _1 (LIST) {
- | repeated int32 element;
+ | repeated int32 array;
| }
|}
""".stripMargin)
@@ -198,7 +198,7 @@ class ParquetSchemaInferenceSuite extends ParquetSchemaTest {
|message root {
| optional group _1 (LIST) {
| repeated group bag {
- | optional int32 element;
+ | optional int32 array_element;
| }
| }
|}
@@ -267,7 +267,7 @@ class ParquetSchemaInferenceSuite extends ParquetSchemaTest {
| optional binary _1 (UTF8);
| optional group _2 (LIST) {
| repeated group bag {
- | optional group element {
+ | optional group array_element {
| required int32 _1;
| required double _2;
| }
@@ -616,7 +616,7 @@ class ParquetSchemaSuite extends ParquetSchemaTest {
"""message root {
| optional group f1 (LIST) {
| repeated group bag {
- | optional int32 element;
+ | optional int32 array_element;
| }
| }
|}
@@ -648,7 +648,7 @@ class ParquetSchemaSuite extends ParquetSchemaTest {
nullable = true))),
"""message root {
| optional group f1 (LIST) {
- | repeated int32 element;
+ | repeated int32 array;
| }
|}
""".stripMargin)
--
GitLab