Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
0.10.1
-
None
-
None
Description
A = LOAD 'job_confs' USING org.apache.hcatalog.pig.HCatLoader();
B = FILTER A by grid == 'cluster1' and dt < '2012_12_01' and dt > '2012_11_20';
C = FILTER B by params#'mapreduce.job.user.name' == 'userx';
D = FOREACH B generate dt, grid, params#'mapreduce.job.user.name' as user,
params#'mapreduce.job.name' as job_name, job_id,
params#'mapreduce.job.cache.files';
dump D;
The query gives the below warning and ends up scanning the whole table instead of pushing the partition key filters grid and dt.
[main] WARN org.apache.pig.newplan.PColFilterExtractor - No partition filter
push down: Internal error while processing any partition filter conditions in
the filter after the load
Works fine if the second filter is on a column with simple datatype like chararray instead of map.