Details
-
Sub-task
-
Status: Open
-
Major
-
Resolution: Unresolved
-
impala 2.5.1
-
None
-
None
Description
Partitions in the Catalog already store the literals of their clustering columns. The tables also already store the clustering column names.
THis information is stored redundantly in partition location strings like "hdfs://db/table/part_col_1=part_val_1/part_col_2=part_val_2/". We could save catalog memory by reducing that redundancy.