sorting - Cassandra sort and a changing clustering key -

i have data modeling question cases data needs sorted keys can modified. , have user table

{    dept_id text,    user_id text,    user_name text,    mod_date timestamp    primary key (dept_id,user_id) } 

now can query cassandra users dept_id.

what if wanted query users in dept, sorted mod_date.

so, 1 way to

{    dept_id text,    mod_date timestamp,    user_id text,    user_name text,    primary key (dept_id, mod_date,user_id) } 

but, mod_date changes every time user name updated. can't part of clustering key.

attempt 1:
don't update row instead create new record every update.

so, record user foo below {'dept_id1',timestamp1','user_id1','foo'} , name changed 'bar' , 'baz' . in case add row table, like

{'dept_id1',timestamp3','user_id1','baz'} {'dept_id1',timestamp2','user_id1','bar'} {'dept_id1',timestamp1','user_id1','foo'} 

now can users in dept, sorted mod_date presents different problem.

the data returned duplicated

.

attempt 2 : add column identify head record linked list

{    dept_id text,    mod_date timestamp,    user_id text,    user_name text,    next_record text    primary key (dept_id,mod_date,user_id) } 

every time update happens adds row , adds pk of new record.

{'dept_id1',timestamp3','user_id1','baz','head'} {'dept_id1',timestamp2','user_id1','bar','dept_id1#timestamp3'} {'dept_id1',timestamp1','user_id1','foo','dept_id1#timestamp2'} 

and add secondary index 'next_record' column.

now can support users in dept, sorted mod_date by

select * users dept_id=':dept' , next_record='head' order mod_date.

but looks involved solution , perhaps missing , simpler solution ..

the other option delete , insert high frequency changes think cassandra has issues tombstones.

suggestions/feedback welcome. !

as see, simplest way sorting users on application (client code) side. use dept partition key, means users in 1 dept can handled 1 cassandra node, there no many users in 1 dept , users can sorted on application side fast enough.