Closed adrian-ionescu closed 8 years ago
The previous solution involved registering a DataFrame with 0 columns as a temporary table and then running a select * from it. This works in spark 1.5.1, but not with 1.6.x.
I raised https://issues.apache.org/jira/browse/SPARK-16329, but we can just avoid doing that altogether by calling DataFrameWriter.insertInto() instead. That's what this patch is about.
ship it
(please make a related issue in github too for this bug, with reference to the mantis issue)
The previous solution involved registering a DataFrame with 0 columns as a temporary table and then running a select * from it. This works in spark 1.5.1, but not with 1.6.x.
I raised https://issues.apache.org/jira/browse/SPARK-16329, but we can just avoid doing that altogether by calling DataFrameWriter.insertInto() instead. That's what this patch is about.