> Disclaimer: Sending this to tez mailing list even though a Hive query >fails, because this query passes on MapReduce but fails on Tez. .. > I have a hive table with two partitions p=1 and p=2. p=1 contains >Sequence Files with key as BytesWritable, p=2 contains Sequence Files >with key as Text. A simple count(*) on the table fails with:
I suspect this is entirely due to the HiveSplitGenerator (which is Tez only), the Tez split grouper and the HiveInputFormat reading from TezGroupedSplits (also, only enabled for Tez). Just have an example table produced by ³export table² and file a HIVE JIRA. This is the exact same issue reported earlier - but this is somewhat of a user error, since this scenario cannot be generated if you purely generate all your tables using Hive inserts. Cheers, Gopal
