Open zsxwing opened 2 years ago
@Kimahriman I saw you submitted a PR previously in #699. But as I mentioned above, we don't want to use Spark's current output format. Are you still interested in working on this with a different format? In order to do this, we would need to parse the SHOW PARTITIONS
command in DeltaSqlParser
to execute our own SHOW PARTITIONS command.
I don't really have a use case for it anymore so don't really have the time to spend on it. Mostly used it before as a way to learn some DataSourceV2 things, it's just very annoying the format that's used for the existing show partitions behavior
cc @JassAbidi - could you tackle this one? Thanks!
Hi, @zsxwing @dennyglee I had some time to fix this issue and prepared delta-io/delta/pull/1667 It introduces not only 'SHOW PARTITIONS' but also an additional feature 'DETAIL'. This is what we need in our company because getting the list of files and other partition details is not available now
Hi, @zsxwing @dennyglee Sorry to bother you again. Do you know who can review https://github.com/delta-io/delta/pull/1667?
It would be great to support
SHOW PARTITIONS
in Delta so that people can check the partition values in a table. Output example,The column names of the output should be the partition column names and each row outputs the partition values for one partition.
Syntax:
Note: we don't want to follow Spark's current SHOW PARTITIONS output format because it's not easy to write code to consume this format. For example, the following example puts two values in one cell.