Is there a way to get a list of columns with non-missing values? Note that I got more then 2000 columns, so check every column manually is not possible.
I tried to use metadata:
ANALYZE TABLE test COMPUTE STATISTICS FOR COLUMNS;
DESCRIBE FORMATTED test;
However this is not really helpful because it does not give me useful information on how the columns are actually filled.
For the beginning, an output table as described in https://cwiki.apache.org/confluence/display/Hive/StatsDev#StatsDev-ExistingTables.
using the command
desc formatted concurrent_delete_different partition(ds='tomorrow') name;
would be enough. Unfortunately, this command doesn't work using Hue.