如何在几天内迭代选择语句

Question

I need to iterate a select statement over a range of days. The purpose is to get the volume of open records in our various groups at a moment in time each day in the past year, by using the open and close time stamps of the records. The select statement is something like this:

select I.INC_GROUP, count(I.INC_GROUP) as INC_VOLUME
from XXX.INCIDENT i
where
I.OPEN_TIME < to_date('2015/09/21 08:00:00', 'YYYY/MM/DD HH24:MI:SS') and
I.CLOSE_TIME > to_date('2015/09/21 08:00:00', 'YYYY/MM/DD HH24:MI:SS')
group by I.INC_GROUP

I am trying to iterate it for 8AM today-366 to 8AM today-1. My output would be inserted into a table and would consist of the INC_GROUP, INC_VOLUME, and the date of the iteration (first record would be 22-SEP-14 08:00:00).

Answer 1

Use a CONNECT BY to generate a set of numbers and use these in conjunction with date logic (DATE + 1 will add a day). This will generate you a row for each day.

SELECT TRUNC(SYSDATE - LEVEL) + (1/24*8)
FROM dual
CONNECT BY LEVEL <= 365

This can then be incorporated to another data set, for example:

WITH days AS (
  SELECT
    TRUNC(SYSDATE - LEVEL) + (1/24*8) day_start
  -- This adds a day to the above and removes a second
  , TRUNC(SYSDATE - LEVEL + 1) + (1/24*8) - (1/24/60/60) day_end    
  FROM dual
  CONNECT BY LEVEL <= 365
)
SELECT 
  days.day_start
, COUNT(yt.date_col) num
FROM days
LEFT JOIN your_table yt ON yt.date_col BETWEEN days.day_start AND days.day_end

This would give you a row for each day and a number of matching rows from your_table that fell between 8am and 7:59:59 the following day.