Poor performance for queries which target multiple separate primary keys

[yugabyte]

Poor performance for queries which target multiple separate primary keys

[kmuthukk]

@nickhristov : thanks for reporting the issue.

One comment, with CQL's CREATE TABLE syntax:
PRIMARY KEY (col_a, col_b)
is effectively same as:
PRIMARY KEY ((col_a), col_b)
and both would result in col_a being hash/partition column, and col_b being a clustering column,

If you wanted both columns be part of the hash/partition columns in the PRIMARY KEY, then you'll need an extra set of parens as below:
PRIMARY KEY ((col_a, col_b))

Nevertheless, in either case, looks like we do have an issue with the IN clause not being optimized to do an index lookup - and it might be that we are doing a scan here instead. We will get a fix out to you soon.

@m-iancu : Assigning this to you.