Partition over date range

Question

Is there a way to have a count be partition over a date range
count(records) over(partition by ids range start_date to end_date)

Answer 1

Hi,
Whenever you have a question, post  a little sample data (CREATE TABLE and INSERT statements, relevant columns only) for all tables involved, and also post the results you want from that data.
Explain, using specific examples, how you get those results from that data.
Always say which version of Oracle you're using (e.g., 11.2.0.2.0).
See the forum FAQ: https://forums.oracle.com/message/9362002
Since I don't have a copy of your table, I'll use the scott.emp table to illustrate.  Of course, there's nothing about cloudy days in scott.emp, so let's count the number of employees in non-magement positions, that is, anyone who has a job other than 'MANAGER' or 'PRESIDENT'.
For every employee hired, we want to see the total number of non-management employees hired so far in the same year, and also so far in the same month, like this:
ENAME      HIREDATE    JOB         YEAR_CNT  MONTH_CNT
SMITH      17-Dec-1980 CLERK              1          1
ALLEN      20-Feb-1981 SALESMAN           1          1
WARD       22-Feb-1981 SALESMAN           2          2
JONES      02-Apr-1981 MANAGER            2          0
BLAKE      01-May-1981 MANAGER            2          0
CLARK      09-Jun-1981 MANAGER            2          0
TURNER     08-Sep-1981 SALESMAN           3          1
MARTIN     28-Sep-1981 SALESMAN           4          2
KING       17-Nov-1981 PRESIDENT          4          0
JAMES      03-Dec-1981 CLERK              6          2
FORD       03-Dec-1981 ANALYST            6          2
MILLER     23-Jan-1982 CLERK              1          1
SCOTT      19-Apr-1987 ANALYST            1          1
ADAMS      23-May-1987 CLERK              2          1
Here's one way to do that:
WITH    got_derived_values   AS
    SELECT  ename, hiredate, job
    ,       CASE
                WHEN  job  NOT IN ('MANAGER', 'PRESIDENT')
                THEN  'NON-MANAGEMENT'
            END                                    AS non_man
    ,       hiredate - TRUNC (hiredate, 'MONTH')   AS month_range
    ,        hiredate - TRUNC (hiredate, 'YEAR' )  AS year_range
    FROM    scott.emp
SELECT    ename, hiredate, job
,         COUNT (non_man) OVER ( ORDER BY  hiredate
                                 RANGE BETWEEN  year_range PRECEDING
                                       AND      CURRENT ROW
                               )  AS year_cnt
,         COUNT (non_man) OVER ( ORDER BY  hiredate
                                 RANGE BETWEEN  month_range PRECEDING
                                       AND      CURRENT ROW
                               )  AS month_cnt
FROM      got_derived_values
ORDER BY  hiredate
You've probably heard this before, but it's worth repeating:
Use DATE (or TIMESTAMP) columns to store information about dates.  Using any other datatype is simply asking for trouble.