[jira] [Created] (CASSANDRA-7630) compound secondary index on primary key fields

["Carlo Bertuccini (JIRA)" <ji...@apache.org>]

Carlo Bertuccini created CASSANDRA-7630:
-------------------------------------------

             Summary: compound secondary index on primary key fields
                 Key: CASSANDRA-7630
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7630
             Project: Cassandra
          Issue Type: Improvement
          Components: Core
            Reporter: Carlo Bertuccini


Hi all,
I was wondering if a secondary index could be composite by more columns which are parts of the primary key.

Example 
{code}
CREATE TABLE users (
  firstname text,
  lastname text,
  email text,
  age int,
  city text,
  state text,
  PRIMARY KEY ((firstname), email, lastname)
)
{code}

This index should now be possible
{code}
CREATE INDEX surname_by_city on (city, lastname);
{code}
Since we surely know that lastname is always populated.

Now the challenge is (and in my mind looks possible!-) to perform such queries:
{code}
select * from users where city = 'seattle';
select * from users where city = 'seattle' and lastname = 'kirkland';
select * from users where firstname='john' and city = 'seattle';
select * from users where firstname='john' and city = 'seattle' and lastname = 'kirkland';
{code}

If this possibility has already been evaluated and discarded/being developed I ask sorry but I couldn't find anywhere.

Regards,
Carlo



--
This message was sent by Atlassian JIRA
(v6.2#6252)