Supports new tpc-ds dsdgen tool about spark-sql-perf

[fraPace]

org.apache.spark.SparkException: Task failed while writing rows.
at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation.org$apache$spark$sql$sources$InsertIntoHadoopFsRelation$$writeRows$1(commands.scala:161)
at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation$$anonfun$insert$1.apply(commands.scala:132)
at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation$$anonfun$insert$1.apply(commands.scala:132)
at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:63)
at org.apache.spark.scheduler.Task.run(Task.scala:70)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:213)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)
Caused by: java.lang.ArrayIndexOutOfBoundsException: 0
at org.apache.spark.sql.catalyst.expressions.GenericRow.apply(rows.scala:77)
at org.apache.spark.sql.catalyst.expressions.BoundReference.eval(BoundAttribute.scala:37)
at org.apache.spark.sql.catalyst.expressions.Cast.eval(Cast.scala:435)
at org.apache.spark.sql.catalyst.expressions.Cast.eval(Cast.scala:435)
at org.apache.spark.sql.catalyst.expressions.Alias.eval(namedExpressions.scala:118)
at org.apache.spark.sql.catalyst.expressions.InterpretedMutableProjection.apply(Projection.scala:68)
at org.apache.spark.sql.catalyst.expressions.InterpretedMutableProjection.apply(Projection.scala:52)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
at scala.collection.Iterator$$anon$13.next(Iterator.scala:372)
at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation.org$apache$spark$sql$sources$InsertIntoHadoopFsRelation$$writeRows$1(commands.scala:153)
... 8 more

[fraPace]

On the workers I have the following:

DBGEN2 Population Generator (Version 1.4.0)
Copyright Transaction Processing Performance Council (TPC) 2001 - 2015
15/09/23 12:39:38 ERROR InsertIntoHadoopFsRelation: Aborting task.
java.lang.ArrayIndexOutOfBoundsException: 0
at org.apache.spark.sql.catalyst.expressions.GenericRow.apply(rows.scala:77)
at org.apache.spark.sql.catalyst.expressions.BoundReference.eval(BoundAttribute.scala:37)
at org.apache.spark.sql.catalyst.expressions.Cast.eval(Cast.scala:435)
at org.apache.spark.sql.catalyst.expressions.Cast.eval(Cast.scala:435)
at org.apache.spark.sql.catalyst.expressions.Alias.eval(namedExpressions.scala:118)
at org.apache.spark.sql.catalyst.expressions.InterpretedMutableProjection.apply(Projection.scala:68)
at org.apache.spark.sql.catalyst.expressions.InterpretedMutableProjection.apply(Projection.scala:52)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
at scala.collection.Iterator$$anon$13.next(Iterator.scala:372)
at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation.org$apache$spark$sql$sources$InsertIntoHadoopFsRelation$$writeRows$1(commands.scala:153)
at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation$$anonfun$insert$1.apply(commands.scala:132)
at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation$$anonfun$insert$1.apply(commands.scala:132)
at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:63)
at org.apache.spark.scheduler.Task.run(Task.scala:70)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:213)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)

[yhuai]

@Nosfe What is the version of dsdgen (when did you download it)? Can you try https://github.com/gregrahn/tpcds-kit?

[fraPace]

@yhuai

Hi, first of all thanks for the reply.
The dsdgen version is 1.4.0, I downloaded it yesterday.

No Exceptions raised if I use the version in the link. This might means that the new version of dsdgen has new fields or just a different schema.
Any plans to support it?

When I have time I will try to fix the problem by myself and open a pull request, but I have no ETA at the moment.

Cheers

[yhuai]

@Nosfe Thank you for trying the repo.

I have not tried the new tool. It will be good to investigate what has been changed. Since it is a standard benchmark, I guess the schemas of tables are not changed. Maybe the problem is that the text format has been changed or the new tool takes different parameters. I'd like to take a look (and also update our tool to generate all tables), but I do not have ETA on that as well. How about we change the title of this issue to something like "Supports new tpc-ds dsdgen tool"? Then, we can share what we find at here.

[fraPace]

Sounds good.
Already made the change in the title.

[fraPace]

@yhuai
I have taken a look at the new dsdgen and compared it with the one from the repo that you linked me. There are a couple of big changes.

1. option -filter is no more, they changed "FILTER" to "_FILTER", i believe using -f can fix that.
2. There is no more STDOUT output. This means that we have to let dsdgen write the data on disk, then read them and convert to the appropriate format.

I will keep you updated and try to fix the code during spare time.