Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hive >> mail # dev >> Question regarding test query archive.q


Copy link to this message
-
Question regarding test query archive.q
Hi,

The query test - archive.q that creates and uses hadoop archives was broken
for hadoop 1.0. I have made the changes appropriately in the shims.
https://issues.apache.org/jira/browse/HIVE-3338

There is one issue however:

For archive.q test, in case of hadoop 20, the golden results seem to expect
different values for the following sequence of queries:

SELECT SUM(hash(col)) FROM (SELECT transform using 'tr "\t" "_"' AS col
FROM (SELECT * FROM old_name WHERE ds='1') subq1) subq2;
ALTER TABLE old_name RENAME TO new_name;
SELECT SUM(hash(col)) FROM (SELECT transform using 'tr "\t" "_"' AS col
FROM (SELECT * FROM new_name WHERE ds='1') subq1) subq2;

The result for the first query is 48656137 whereas after a simple rename of
the table, the golden results expect NULL. This does not make sense to me.
With my changes for 1.0, I see that the expected value of 48656137 is
returned in both cases. I would be very grateful if someone can explain the
reasoning behind expecting NULL in the case of hadoop 20.

Thanks,
Vikram.
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB