Skip to content

[SPARK-32380][SQL] fixed spark3.0 access hive table while data in hbase problem #29178

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
wants to merge 1 commit into from
Closed

[SPARK-32380][SQL] fixed spark3.0 access hive table while data in hbase problem #29178

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
4 changes: 3 additions & 1 deletion sql/hive/src/main/scala/org/apache/spark/sql/hive/TableReader.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -299,7 +299,9 @@ class HadoopTableReader(
*/
private def createHadoopRDD(localTableDesc: TableDesc, inputPathStr: String): RDD[Writable] = {
val inputFormatClazz = localTableDesc.getInputFileFormatClass
if (classOf[newInputClass[_, _]].isAssignableFrom(inputFormatClazz)) {
if (!inputFormatClazz.getName.
equalsIgnoreCase("org.apache.hadoop.hive.hbase.HiveHBaseTableInputFormat")
Copy link
Member

@HyukjinKwon HyukjinKwon Jul 22, 2020

Choose a reason for hiding this comment

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

Do you know why org.apache.hadoop.hive.hbase.HiveHBaseTableInputFormat implements new Hadoop inputformat interface but doesn't work?

Copy link

@yangBottle yangBottle Jan 8, 2021

Choose a reason for hiding this comment

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

It looks like the new MapReduce API (org.apache.hadoop.mapreduce) used when creating NewHadoopRDD , but The getsplits method of HiveHBaseTableInputFormat is implemented by org.apache.hadoop.mapred API，so some initialization operations (Table、connection) are not done,so the obtained variable table is null.And when using methord createOldHadoopRDD will use the org.apache.hadoop.mapred API,and some initialization operations (Table、connection) are doing,so it can work well.

&& classOf[newInputClass[_, _]].isAssignableFrom(inputFormatClazz)) {
Copy link
Member

@dongjoon-hyun dongjoon-hyun Jul 21, 2020

Choose a reason for hiding this comment

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

Do you think we can have a test case for this?

Copy link
Author

@DeyinZhong DeyinZhong Jul 22, 2020

Choose a reason for hiding this comment

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

I try Spark2.4.3 can work well.
spark3.0.0 will createNewHadoopRDD, and call 'val allRowSplits = inputFormat.getSplits(new JobContextImpl(_conf, jobId)).asScala' in method getPartitions, this will call TableInputFormatBase.getSplits(JobContext context), but the variable table is null, so throw the execption;

createNewHadoopRDD(localTableDesc, inputPathStr)
} else {
createOldHadoopRDD(localTableDesc, inputPathStr)
Expand Down