spark git commit: [SPARK-7447] [SQL] Don't re-merge Parquet schema when the relation is deserialized
Repository: spark
Updated Branches:
refs/heads/branch-1.4 0ed376afa -> 898be6248
[SPARK-7447] [SQL] Don't re-merge Parquet schema when the relation is
deserialized
JIRA: https://issues.apache.org/jira/browse/SPARK-7447
`MetadataCache` in `ParquetRelation2` is annotated as `transient`. When
`ParquetRelation2` is deserialized, we ask `MetadataCache` to refresh and
perform schema merging again. It is time-consuming especially for very many
parquet files.
With the new `FSBasedParquetRelation`, although `MetadataCache` is not
`transient` now, `MetadataCache.refresh()` still performs schema merging again
when the relation is deserialized.
Author: Liang-Chi Hsieh
Closes #6012 from viirya/without_remerge_schema and squashes the following
commits:
2663957 [Liang-Chi Hsieh] Merge remote-tracking branch 'upstream/master' into
without_remerge_schema
6ac7d93 [Liang-Chi Hsieh] Merge remote-tracking branch 'upstream/master' into
without_remerge_schema
b0fc09b [Liang-Chi Hsieh] Don't generate and merge parquetSchema multiple times.
(cherry picked from commit 339905578790fa37fcad9684b859b443313a5aa2)
Signed-off-by: Cheng Lian
Project: http://git-wip-us.apache.org/repos/asf/spark/repo
Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/898be624
Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/898be624
Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/898be624
Branch: refs/heads/branch-1.4
Commit: 898be6248940acddab6f40af0fbc7e7abb3adb76
Parents: 0ed376a
Author: Liang-Chi Hsieh
Authored: Sun May 17 15:42:21 2015 +0800
Committer: Cheng Lian
Committed: Sun May 17 15:42:40 2015 +0800
--
.../apache/spark/sql/parquet/newParquet.scala | 32 +++-
1 file changed, 18 insertions(+), 14 deletions(-)
--
http://git-wip-us.apache.org/repos/asf/spark/blob/898be624/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
--
diff --git
a/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
b/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
index 946062f..bcbdb1e 100644
--- a/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
+++ b/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
@@ -340,7 +340,7 @@ private[sql] class ParquetRelation2(
// Schema of the actual Parquet files, without partition columns
discovered from partition
// directory paths.
-var dataSchema: StructType = _
+var dataSchema: StructType = null
// Schema of the whole table, including partition columns.
var schema: StructType = _
@@ -379,19 +379,23 @@ private[sql] class ParquetRelation2(
f -> new Footer(f.getPath, parquetMetadata)
}.seq.toMap
- dataSchema = {
-val dataSchema0 =
- maybeDataSchema
-.orElse(readSchema())
-.orElse(maybeMetastoreSchema)
-.getOrElse(sys.error("Failed to get the schema."))
-
-// If this Parquet relation is converted from a Hive Metastore table,
must reconcile case
-// case insensitivity issue and possible schema mismatch (probably
caused by schema
-// evolution).
-maybeMetastoreSchema
- .map(ParquetRelation2.mergeMetastoreParquetSchema(_, dataSchema0))
- .getOrElse(dataSchema0)
+ // If we already get the schema, don't need to re-compute it since the
schema merging is
+ // time-consuming.
+ if (dataSchema == null) {
+dataSchema = {
+ val dataSchema0 =
+maybeDataSchema
+ .orElse(readSchema())
+ .orElse(maybeMetastoreSchema)
+ .getOrElse(sys.error("Failed to get the schema."))
+
+ // If this Parquet relation is converted from a Hive Metastore
table, must reconcile case
+ // case insensitivity issue and possible schema mismatch (probably
caused by schema
+ // evolution).
+ maybeMetastoreSchema
+.map(ParquetRelation2.mergeMetastoreParquetSchema(_, dataSchema0))
+.getOrElse(dataSchema0)
+}
}
}
-
To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org
For additional commands, e-mail: commits-h...@spark.apache.org
spark git commit: [SPARK-7447] [SQL] Don't re-merge Parquet schema when the relation is deserialized
Repository: spark
Updated Branches:
refs/heads/master edf09ea1b -> 339905578
[SPARK-7447] [SQL] Don't re-merge Parquet schema when the relation is
deserialized
JIRA: https://issues.apache.org/jira/browse/SPARK-7447
`MetadataCache` in `ParquetRelation2` is annotated as `transient`. When
`ParquetRelation2` is deserialized, we ask `MetadataCache` to refresh and
perform schema merging again. It is time-consuming especially for very many
parquet files.
With the new `FSBasedParquetRelation`, although `MetadataCache` is not
`transient` now, `MetadataCache.refresh()` still performs schema merging again
when the relation is deserialized.
Author: Liang-Chi Hsieh
Closes #6012 from viirya/without_remerge_schema and squashes the following
commits:
2663957 [Liang-Chi Hsieh] Merge remote-tracking branch 'upstream/master' into
without_remerge_schema
6ac7d93 [Liang-Chi Hsieh] Merge remote-tracking branch 'upstream/master' into
without_remerge_schema
b0fc09b [Liang-Chi Hsieh] Don't generate and merge parquetSchema multiple times.
Project: http://git-wip-us.apache.org/repos/asf/spark/repo
Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/33990557
Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/33990557
Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/33990557
Branch: refs/heads/master
Commit: 339905578790fa37fcad9684b859b443313a5aa2
Parents: edf09ea
Author: Liang-Chi Hsieh
Authored: Sun May 17 15:42:21 2015 +0800
Committer: Cheng Lian
Committed: Sun May 17 15:42:21 2015 +0800
--
.../apache/spark/sql/parquet/newParquet.scala | 32 +++-
1 file changed, 18 insertions(+), 14 deletions(-)
--
http://git-wip-us.apache.org/repos/asf/spark/blob/33990557/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
--
diff --git
a/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
b/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
index 946062f..bcbdb1e 100644
--- a/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
+++ b/sql/core/src/main/scala/org/apache/spark/sql/parquet/newParquet.scala
@@ -340,7 +340,7 @@ private[sql] class ParquetRelation2(
// Schema of the actual Parquet files, without partition columns
discovered from partition
// directory paths.
-var dataSchema: StructType = _
+var dataSchema: StructType = null
// Schema of the whole table, including partition columns.
var schema: StructType = _
@@ -379,19 +379,23 @@ private[sql] class ParquetRelation2(
f -> new Footer(f.getPath, parquetMetadata)
}.seq.toMap
- dataSchema = {
-val dataSchema0 =
- maybeDataSchema
-.orElse(readSchema())
-.orElse(maybeMetastoreSchema)
-.getOrElse(sys.error("Failed to get the schema."))
-
-// If this Parquet relation is converted from a Hive Metastore table,
must reconcile case
-// case insensitivity issue and possible schema mismatch (probably
caused by schema
-// evolution).
-maybeMetastoreSchema
- .map(ParquetRelation2.mergeMetastoreParquetSchema(_, dataSchema0))
- .getOrElse(dataSchema0)
+ // If we already get the schema, don't need to re-compute it since the
schema merging is
+ // time-consuming.
+ if (dataSchema == null) {
+dataSchema = {
+ val dataSchema0 =
+maybeDataSchema
+ .orElse(readSchema())
+ .orElse(maybeMetastoreSchema)
+ .getOrElse(sys.error("Failed to get the schema."))
+
+ // If this Parquet relation is converted from a Hive Metastore
table, must reconcile case
+ // case insensitivity issue and possible schema mismatch (probably
caused by schema
+ // evolution).
+ maybeMetastoreSchema
+.map(ParquetRelation2.mergeMetastoreParquetSchema(_, dataSchema0))
+.getOrElse(dataSchema0)
+}
}
}
-
To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org
For additional commands, e-mail: commits-h...@spark.apache.org
