Skip to content

[SPARK-29869][SQL] improve error message in HiveMetastoreCatalog#convertToLogicalRelation #26499

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
LantaoJin wants to merge 10 commits into from
Closed

[SPARK-29869][SQL] improve error message in HiveMetastoreCatalog#convertToLogicalRelation #26499

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
dd1efc8
[SPARK-29869][SQL] Make error message more meaningful when HiveMetast…
LantaoJin Nov 13, 2019
1071b50
add suggestion
LantaoJin Nov 13, 2019
1c98c38
add ut
LantaoJin Nov 14, 2019
363afb6
fix
LantaoJin Nov 14, 2019
46c16c0
fix
LantaoJin Nov 14, 2019
1df7b8d
fix ut
LantaoJin Nov 15, 2019
9f29abc
address comments
LantaoJin Nov 15, 2019
ad71656
remove useless imports
LantaoJin Nov 15, 2019
2efaafc
address comments
LantaoJin Nov 19, 2019
3288ffd
revert sameType
LantaoJin Nov 19, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 9 additions & 6 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/DataSource.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -330,7 +330,9 @@ case class DataSource(
* is considered as a non-streaming file based data source. Since we know
* that files already exist, we don't need to check them again.
*/
def resolveRelation(checkFilesExist: Boolean = true): BaseRelation = {
def resolveRelation(
checkFilesExist: Boolean = true,
needPartitionInferring: Boolean = true): BaseRelation = {
val relation = (providingInstance(), userSpecifiedSchema) match {
// TODO: Throw when too much is given.
case (dataSource: SchemaRelationProvider, Some(schema)) =>
Expand Down Expand Up @@ -391,7 +393,7 @@ case class DataSource(
} else {
val globbedPaths = checkAndGlobPathIfNecessary(
checkEmptyGlobPath = true, checkFilesExist = checkFilesExist)
val index = createInMemoryFileIndex(globbedPaths)
val index = createInMemoryFileIndex(globbedPaths, needPartitionInferring)
val (resultDataSchema, resultPartitionSchema) =
getOrInferFileFormatSchema(format, () => index)
(index, resultDataSchema, resultPartitionSchema)
Expand Down Expand Up @@ -427,7 +429,6 @@ case class DataSource(
"in the data schema",
equality)
}

relation
}

Expand Down Expand Up @@ -551,10 +552,12 @@ case class DataSource(
}

/** Returns an [[InMemoryFileIndex]] that can be used to get partition schema and file list. */
private def createInMemoryFileIndex(globbedPaths: Seq[Path]): InMemoryFileIndex = {
private def createInMemoryFileIndex(
globbedPaths: Seq[Path],
needPartitionInferring: Boolean = true): InMemoryFileIndex = {
val fileStatusCache = FileStatusCache.getOrCreate(sparkSession)
new InMemoryFileIndex(
sparkSession, globbedPaths, options, userSpecifiedSchema, fileStatusCache)
new InMemoryFileIndex(sparkSession, globbedPaths, options, userSpecifiedSchema,
fileStatusCache, needPartitionInferring)
}

/**
Expand Down
9 changes: 7 additions & 2 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/InMemoryFileIndex.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -50,7 +50,8 @@ class InMemoryFileIndex(
rootPathsSpecified: Seq[Path],
parameters: Map[String, String],
userSpecifiedSchema: Option[StructType],
fileStatusCache: FileStatusCache = NoopCache)
fileStatusCache: FileStatusCache = NoopCache,
needPartitionInferring: Boolean = true)
extends PartitioningAwareFileIndex(
sparkSession, parameters, userSpecifiedSchema, fileStatusCache) {

Expand All @@ -69,7 +70,11 @@ class InMemoryFileIndex(

override def partitionSpec(): PartitionSpec = {
if (cachedPartitionSpec == null) {
cachedPartitionSpec = inferPartitioning()
cachedPartitionSpec = if (needPartitionInferring) {
inferPartitioning()
} else {
PartitionSpec.emptySpec
}
}
logTrace(s"Partition spec: $cachedPartitionSpec")
cachedPartitionSpec
Expand Down
2 changes: 1 addition & 1 deletion sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveMetastoreCatalog.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -245,7 +245,7 @@ private[hive] class HiveMetastoreCatalog(sparkSession: SparkSession) extends Log
userSpecifiedSchema = Option(updatedTable.dataSchema),
bucketSpec = None,
options = options,
className = fileType).resolveRelation(),
className = fileType).resolveRelation(needPartitionInferring = false),
table = updatedTable)

catalogProxy.cacheTable(tableIdentifier, created)
Expand Down
27 changes: 9 additions & 18 deletions sql/hive/src/test/scala/org/apache/spark/sql/hive/HiveMetastoreCatalogSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -362,29 +362,20 @@ class DataSourceWithHiveMetastoreCatalogSuite
}
}

test("SPARK-29869: HiveMetastoreCatalog#convertToLogicalRelation throws AssertionError") {
test("SPARK-29869: fix HiveMetastoreCatalog#convertToLogicalRelation throws AssertionError") {
withTempPath(dir => {
val baseDir = s"${dir.getCanonicalFile.toURI.toString}/test"
Copy link
Member

@wangyum wangyum Nov 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

test -> non_partition_table ?

Copy link
Contributor Author

@LantaoJin LantaoJin Nov 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the location can be set to anywhere. test here is fine. non_partition_table will over 100 chars :)

val partitionDir = s"${dir.getCanonicalFile.toURI.toString}/test/dt=20191113"
val file = new Path(partitionDir, "file.parquet")
val fs = file.getFileSystem(new Configuration())
fs.createNewFile(file)
withTable("test") {
withSQLConf(HiveUtils.CONVERT_METASTORE_PARQUET.key -> "true",
SQLConf.PARALLEL_PARTITION_DISCOVERY_THRESHOLD.key -> "0") {
val partitionLikeDir = s"${dir.getCanonicalFile.toURI.toString}/test/dt=20191113"
Copy link
Member

@wangyum wangyum Nov 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

val partitionLikeDir = s"${dir.getCanonicalFile.toURI.toString}/test/dt=20191113" -> val partitionLikeDir = s"$baseDir/dt=20191113"?

spark.range(3).selectExpr("id").write.parquet(partitionLikeDir)
withTable("non_partition_table") {
withSQLConf(HiveUtils.CONVERT_METASTORE_PARQUET.key -> "true") {
spark.sql(
s"""
|CREATE EXTERNAL TABLE `test`(key string)
|ROW FORMAT SERDE
| 'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
|STORED AS INPUTFORMAT
| 'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
|OUTPUTFORMAT
| 'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
|LOCATION '$baseDir'
|CREATE TABLE non_partition_table (id bigint)
Copy link
Contributor

@cloud-fan cloud-fan Nov 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Isn't it a malformed table? Does hive ignore the directories for non-partitioned tables?

Copy link
Contributor Author

@LantaoJin LantaoJin Nov 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems Hive return none when query this table(1.2.1):

hive> select * from xxxxx.xxxx;
OK
Time taken: 25.301 seconds

But no assertion error

Copy link
Contributor Author

@LantaoJin LantaoJin Nov 14, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe you are right. Actually the table LOCATION is /path/tablename/dt=yyyymmdd, But its data file paths are /path/tablename/dt=yyyymmdd/dt=yyyymmdd/xxx.parquet. I guess Hive does not recursively lookup load the data. So it return empty but not error.
And I found if when enable recursively lookup by .option("recursiveFileLookup", true), the inferPartitioning will be disable. So dt=yyyymmdd won't be treated as partitionSpec.

So should I revert the code changes and only keep the assert detail information? Or throws exception instead of assertion, and catch it then rollback to do not use built-in Parquet reader to read?

|STORED AS PARQUET LOCATION '$baseDir'
|""".stripMargin)
val e = intercept[AssertionError](spark.sql("select * from test")).getMessage
assert(e.contains("assertion failed"))
assert(spark.sql("select * from non_partition_table").collect() ===
Array(Row(0), Row(1), Row(2)))
Copy link
Member

@wangyum wangyum Nov 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

assert(spark.sql("select * from non_partition_table").collect() === Array(Row(0), Row(1), Row(2))) -> checkAnswer(spark.table("non_partition_table"), Row(0, 1, 2))?

Copy link
Contributor Author

@LantaoJin LantaoJin Nov 14, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok. I wanna wait more comments to update in one.

}
}
})
Expand Down