I just recently updated to trunk, was lagging a few months behind. Now I'm getting errors like: "Filtering is supported only on partition keys of type string"
It seems some type checking was added on org.apache.hadoop.hive.metastore.parser.ExpressionTree.java:161 which makes sure partition keys are string but my current schema actually contains partition keys which are numbers.
I've removed this check and all is running fine again but was this done for a reason? I think I should be able to partition on anything right?