Details
-
New Feature
-
Status: Closed
-
Major
-
Resolution: Won't Do
-
None
-
None
-
None
Description
Data can sometimes be stored in HDFS in a time-partitioned manner, e.g. /some/input/yyyy/mm/dd. You may want to process this data incrementally, where the output has a format like /some/output/yyyy/mm/dd. It be useful if there is a UDF that handles the incremental processing for you.