Details
Normal Description
First, repro case:
cqlsh:test> create table test ( pk1 text, pk2 timeuuid, data1 text static, data2 text, PRIMARY KEY( pk1, pk2 ) ); cqlsh:test> update test set data1 = 'data1', data2 = 'data2' where pk1 = 'pk1' and pk2 = now(); cqlsh:test> update test set data1 = 'data1', data2 = 'data2' where pk1 = 'pk1' and pk2 = now(); cqlsh:test> select * from test limit 1; pk1 | pk2 | data1 | data2 -----+------+-------+------- pk1 | null | data1 | null (1 rows) cqlsh:test> select * from test limit 2; pk1 | pk2 | data1 | data2 -----+--------------------------------------+-------+------- pk1 | 9b068ee0-b8b0-11e3-a345-49baa9ac32e6 | data1 | data2 (1 rows) cqlsh:test> select * from test limit 3; pk1 | pk2 | data1 | data2 -----+--------------------------------------+-------+------- pk1 | 9b068ee0-b8b0-11e3-a345-49baa9ac32e6 | data1 | data2 pk1 | 0af67a40-b8ba-11e3-a345-49baa9ac32e6 | data1 | data2 (2 rows)
I think that: 1) if this is a static columns feature it should be documented so I can use it safely or 2) it should be fixed (return 2 rows with limit 2 for query above).