[
https://issues.apache.org/jira/browse/SPARK-8632?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15209743#comment-15209743
]
Bijay Kumar Pathak edited comment on SPARK-8632 at 3/24/16 4:47 AM:
--------------------------------------------------------------------
I am still having this issue in Spark 1.6.0 on EMR. The job fails with OOM. I
have DataFrame with 250 columns and I am applying UDF on more than 50 of the
columns. I am registering the DataFrame as temptable and applying the UDF in
hive_context sql statement. I am applying the UDF after sort merge join of two
DataFrame (each of around 4GB) and multiple broadcast joins of 22 Dim table.
Below is how I am applying the UDF.
{code:borderStyle=solid}
data_frame.registerTempTable("temp_table")
new_df = hive_context.sql("select python_udf(column_1),python_udf(column_2),
... , from temp_table")
{code}
was (Author: bijay697):
I am still having this issue in Spark 1.6.0 on EMR. The job fails with OOM. I
have DataFrame with 250 columns and I am applying UDF on more than 50 of the
columns. I am registering the DataFrame as temptable and applying the UDF in
hive_context sql statement. I am applying the UDF after sort merge join of two
DataFrame (each of around 4GB) and multiple broadcast joins of 22 Dim table.
Below is how I am applying the UDF.
{code:title=Bar.python|borderStyle=solid}
data_frame.registerTempTable("temp_table")
new_df = hive_context.sql("select python_udf(column_1),python_udf(column_2),
... , from temp_table")
{code}
> Poor Python UDF performance because of RDD caching
> --------------------------------------------------
>
> Key: SPARK-8632
> URL: https://issues.apache.org/jira/browse/SPARK-8632
> Project: Spark
> Issue Type: Bug
> Components: PySpark, SQL
> Affects Versions: 1.4.0
> Reporter: Justin Uang
> Assignee: Davies Liu
> Priority: Blocker
> Fix For: 1.5.1, 1.6.0
>
>
> {quote}
> We have been running into performance problems using Python UDFs with
> DataFrames at large scale.
> From the implementation of BatchPythonEvaluation, it looks like the goal was
> to reuse the PythonRDD code. It caches the entire child RDD so that it can do
> two passes over the data. One to give to the PythonRDD, then one to join the
> python lambda results with the original row (which may have java objects that
> should be passed through).
> In addition, it caches all the columns, even the ones that don't need to be
> processed by the Python UDF. In the cases I was working with, I had a 500
> column table, and i wanted to use a python UDF for one column, and it ended
> up caching all 500 columns.
> {quote}
> http://apache-spark-developers-list.1001551.n3.nabble.com/Python-UDF-performance-at-large-scale-td12843.html
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
