x = 2 AND upper(y)='YUM', the storage handler might be able to handle x = 2 but leave the "residual" upper(y)='YUM' for Hive to deal with. The breakdown need not be non-overlapping; for example, given the predicate x LIKE 'a%b', the storage handler might be able to evaluate the prefix search x LIKE 'a%', leaving x LIKE '%b' as the residual.
@param jobConf contains a job configuration matching the one thatwill later be passed to getRecordReader and getSplits
@param deserializer deserializer which will be used whenfetching rows
@param predicate predicate to be decomposed
@return decomposed form of predicate, or null if no pushdown ispossible at all
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |