Secondary indexes are not a recommended best practice. Better choices: 1. The new 3.0 materialized views. 2. Manual query tables 3. DSE Search/Solr 4a. Stratio Lucene indexing 4b. TupleJump Stargate Lucene indexing
In any case, low cardinality is not recommended for indexes. The goal is that a query on the index will return a modest to moderate number of rows from a single node. Very high cardinality is also not recommended since indexes are local, so fanout is required if the coordinator node cannot supply the desired row limit. -- Jack Krupansky On Thu, Oct 8, 2015 at 5:37 AM, anuja jain <anujaja...@gmail.com> wrote: > I have two questions, > 1. Does creating secondary index on low cardinality columns like of > boolean type helps in read performance any ways? Because there will be only > two values( true and false ) for that column in index table. > > 2. Should secondary indexes be created on clustering columns even if these > columns are not frequently used in where clause of a query? >
