Question

I am trying out some examples with analytical functions and i have created a sql fiddle to understand a count distinct over partition by clause.This is my sqlfiddle.

create table dummy (value1 varchar2(10),value2 varchar2(10));

insert into dummy values ('abc','abc1');
insert into dummy values ('abc','abc1');
insert into dummy values ('abc','abc2');
insert into dummy values ('def','abc1');
insert into dummy values ('ghi','abc2');
insert into dummy values ('xyz','abc3');
insert into dummy values ('xyz','abc3');

select value1,
       value2,
       count(distinct value2) over (partition by value1) as ValCount
from dummy

If you look at the result set, i would expect valcount as 1 for the third row but instead its 2 and am not sure why that’s the case.

Answer 1

abc (value 1) has only 2 distinct values on the second column (abc1 and abc2), and since you count the distinct values on column2 partitioned over column 1 you should indeed get 2