Counting only some rows without where or subqueries?

Question

Is it possible with oracle to count only some of the rows? I mean without excluding the rows in where or using sub queries.
This does not work but I think it illustrates what I want:
SELECT count(flags), count(flags like '%P%'), count(flags like '%F%') FROM table WHERE some_column = 2;
Where the first would return the number of flags rows, while the second contet the numbers that have P in them and the third counted the number with F.
Table would be
Flags Some_column
'AFP' 2
'A* 1
'PF' 1
'P' 2
And so on.
In postgres you can do: count(substring(flags from 'R')), to achieve the above mentioned idea. Does oracle have a way of doing the same?

Answer 1

Your requirement is not clear. But.. Is this what you are trying to achieve?
SQL> with t
  2  as
  3  (select 'AFP' st from dual
  4  union all
  5  select 'A*' from dual
  6  union all
  7  select 'PF'  from dual
  8  union all
  9  select 'P'  from dual
10  )
11  select sum(1) sum1,
12  sum(case when st like '%P%' then 1 else 0 end) sum2,
13  sum(case when st like '%F%' then 1 else 0 end) sum3
14  from t
15  /
      SUM1       SUM2       SUM3
         4          3          2