Narine Kokhlikyan created SPARK-12629:
-----------------------------------------
Summary: SparkR: DataFrame's saveAsTable method has issues with
the signature and HiveContext
Key: SPARK-12629
URL: https://issues.apache.org/jira/browse/SPARK-12629
Project: Spark
Issue Type: Bug
Components: SparkR
Reporter: Narine Kokhlikyan
There are several issues with the saveAsTable method in SparkR. Here is a
summary of some of them. Hope this will help to fix the issues.
1. According to SparkR's saveAsTable(...) documentation, we can call the
"saveAsTable(df, "myfile")" in order to store the dataframe.
However, this signature isn't working. It seems that "source" and "mode" are
forced according to signature.
2. Within the method saveAsTable(...) it tries to retrieve the SQL context and
tries to create/initialize source as parquet, but this is also failing because
the context has to be hiveContext. Based on the error messages I see.
3. In general the method fails when I try to call it with sqlContext
4. Also, it seems that SQL DataFrame.saveAsTable is deprecated, we could use
df.write.saveAsTable(...) instead ...
[~shivaram] [~sunrui] [~felixcheung]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
