PL/SQL - counting values' occurences in a collection - how to optimize?

Question

Hi all,
I'm new here and working currently on my PhD thesis, using Oracle (XE) to do some data manipulation stuff.
Following is defined:
TYPE stat_col_type IS TABLE OF INTEGER;
rows_lens stat_col_type;
Using MULTISET UNION operations, I get the final rows_lens as a set of values, e.g. {1,2,3,1,1,2,2,2,3,3,3,3,3,3,3,1,1,2,1,1,4,5,6,2,2,2,3,3}.
What I need is to count the occurences of the unique values :
value of 1 : 7 occurences
value of 2 : 8 occurences
value of 3 : 10 occurences
value of 4 : 1 occurence
value of 5 : 1 occurence
value of 6 : 1 occurence
Ideally, the result should be a separate collection where the values become indexes (there are numbers, natural, unique and countable) and the number of occurences - their values :
output_col( value ) := number-of-occurences ;
Of course, I could just go through all the elements incrementing the counters in the output collection, but surely there must be a most efficient way - probably using POWERMULTISET or some kind of SUBSETS. As haven't been using /developing in/ Oracle for a longer period... it's better to ask :)
Thanks in advance for your suggestions.
Regards,
Bart
Edited by: user3698166 on 2010-01-16 13:44

Answer 1

What I need is to count the occurences of the unique values :
Following is defined:Why don't you define in PLSQL and not in SQL?
It would be so much easier:
SQL> create type stat_col_type as table of integer
Type created.
SQL> select column_value, count (column_value) cnt
  from table (
          stat_col_type (1,
                          2,
                          3,
                          1,
                          1,
                          2,
                          2,
                          2,
                          3,
                          3,
                          3,
                          3,
                          3,
                          3,
                          3,
                          1,
                          1,
                          2,
                          1,
                          1,
                          4,
                          5,
                          6,
                          2,
                          2,
                          2,
                          3,
                          3
group by column_value
order by column_value
COLUMN_VALUE        CNT
           1          7
           2          8
           3         10
           4          1
           5          1
           6          1
6 rows selected.
SQL> drop type stat_col_type
Type dropped.