SQL Hadoop Hive vs Toad for ZC890515F1055143925AE4FB85B86EC0

Question

I am trying to construct a query for Toad, however, the following is not working.

select count (distinct t.column1, t.column2)
from schema.table
;

However, the above query works just fine in Hadoop Hive. Any suggestions on refining the query so it works for Toad?

Answer 1

Try to concatenate them:

select count(distinct concat(t.column1, t.column2))
from schema.table t

Answer 2

You can use a subquery:

select count(1)
from (select distinct t.column1, t.column2 from schema.table) as t1
;

Answer 3

Emulating the behavior is a little tricky. The safest method is probably:

select sum(case when seqnum = 1 and column1 is not null and column2 is not null then 1 else 0 end)
from (select t.*,
             row_number() over (partition by column1, column2 order by column1) as seqnum
      from t
     ) t

(The order by column doesn't matter. Many databases require one so I regularly include it.)

This version works for any database, not just DB2.

The issue is that Hive does not count a row if any of the values are NULL , which this takes into account.

The use of select distinct in a subquery is close, but it counts NULL values -- and that change may not be appropriate for other columns in your query.

Concatenating columns together comes closer. However you have problems when there are overlapping values (say '12'/'3' and '1'/'23').