Group by on date field with time difference of up to 1 minute

Question

Hi,
I need to do a group by on a date field and some other fields, but want to include in the same group rows that are identical but may have a difference of up to a minute in the date field.
group by trunc(datefield,'MI') wouldn't work in cases like 00:00:50 and 00:01:10 (20 sec difference).
I can do a semi join with where exists and between on the date field, but am hoping for another solution.
any ideas?
thanks.
Edited by: Pyrocks on Sep 14, 2011 2:16 PM

Answer 1

Hi,
It sounds like you want something like this:
WITH     got_new_grp     AS
     SELECT     x.*     -- or whatever columns you need
     ,     CASE
              WHEN  datefield > MAX (datefield)
                         OVER ( ORDER BY      datefield
                                ROWS BETWEEN  UNBOUNDED PRECEDING
                                     AND          1         PRECEDING
                              ) + ( 1 / (24 * 60))
              THEN  1
          END     AS NEW_GRP
     FROM     table_x     x
--     WHERE     ...
,     got_grp          AS
     SELECT     n.*
     ,     COUNT (new_group) OVER (ORDER BY  datefield)     AS grp
     FROM     got_new_grp     n
SELECT       MIN (datefield)     AS start_datefield
,       MAX (datefield)     AS end_datefield
,       COUNT (*)          AS cnt
FROM       got_grp
GROUP BY  grp
ORDER BY  grp
;Of course, this makes assumptions for all the things you didn't explain.
I hope this answers your question.
If not, post a little sample data (CREATE TABLE and INSERT statements, relevant columns only), and the results you want from that data.
Explain, using specific examples, how you get those results from that data. Be sure to include an example like the one Kim posted, where rows a and c are more than a minute apart, but both are within a minute of row b, or are connected by a chain of rows each less than 1 minute apart from its neighbor.
Always say what version of Oracle you're using.