Support for range partitioning in USQL
Lacking this feature, I have to create distinct partitions for all the dates plus it does not support high precision datetime datatypes. I ran into a similar limitation with HIVEQL on HDINSIGHT clusters.
If you use the same column for partitioning, it overwrites the column data.
I have a 700 million row table which results in 3423 distinct partitions.
Trying to create these partitions @pdate1.......@pdate3423 results in a gigantic script.
The USQL job is stuck on "preparing" phase for 3 hrs and then quits without doing anything.
DESCRIPTION Job was cancelled by Yarn.
MESSAGE Job was cancelled.