Grouping totals by date ranges

Question

Hi All,
I have an ASP.NET application, in which I have to get totals from a table using different criteria, which I do like this:
SELECT DISTINCT
SUM(CASE WHEN MYCONDITION1 THEN 1 ELSE 0 END) AS TOTAL1,
SUM(CASE WHEN MYCONDITION2 THEN 1 ELSE 0 END) AS TOTAL2
FROM TABLE1, TABLE2
WHERE COMMON_CONDITION1 AND COMMON_CONDITION2
AND BETWEEN DATE1 AND DATE2;
This works fine and I get the intended result.
Now, I have to repeat this for every week for the last 12 months, excluding holidays period. So, I generate a set of date ranges which will be used in the queries. So, I repeat the above sql statement for all the date ranges, which is a lengthy process.
I have to get the totals for every week. For example from 26-Sep-2010 to 02-Oct-2010, 19-Sep-2010 to 25-Sep-2010, 12-Sep-2010 to 18-Sep-2010, etc.. How should I put those ranges in the group by and select statement, because, I don't have them in a table.
How can I do that in a single shot and get all totals for each date range. Please help me with sql.
Thank you.
Edited by: 800729 on Oct 7, 2010 4:33 AM

Answer 1

Hi,
Welcome to the forum!
800729 wrote:
Hi All,
I have an ASP.NET application, in which I have to get totals from a table using different criteria, which I do like this:
SELECT DISTINCT
SUM(CASE WHEN MYCONDITION1 THEN 1 ELSE 0 END) AS TOTAL1,
SUM(CASE WHEN MYCONDITION2 THEN 1 ELSE 0 END) AS TOTAL2
FROM TABLE1, TABLE2
WHERE COMMON_CONDITION1 AND COMMON_CONDITION2
AND BETWEEN DATE1 AND DATE2;Does that really work? Don't you have to do something like
AND     date_column  BETWEEN DATE1 AND DATE2Post actual code, simplified if necessary, but still something that you can run.
This works fine and I get the intended result.
Now, I have to repeat this for every week for the last 12 months, excluding holidays period. So, I generate a set of date ranges which will be used in the queries. So, I repeat the above sql statement for all the date ranges, which is a lengthy process.
I have to get the totals for every week. For example from 26-Sep-2010 to 02-Oct-2010, 19-Sep-2010 to 25-Sep-2010, 12-Sep-2010 to 18-Sep-2010, etc.. How should I put those ranges in the group by and select statement, because, I don't have them in a table.Are you saying your current query produces 1 row, but you want something that has 53 rows, one for each week in the last year?
If so, add a "GROUP BY x" clause, as Sabnkar suggested, where x is a distinct value for each week.
You will probably *not" want to use TO_CHAR (date_column, 'WW'), especially if all of the output is not in the same calendar year.
If your weeks always begin on Sunday, then TRUNC (date_column, 'D') would be a better GROUP BY exptression, but it depeneds on your NLS settings.
GROUP BY TRUNC (date_column + 1, 'IW') - 1 will do the same thing, regardless of your NLS settings.
You may not need a table tha has one row per week, but if you do, Sauhik showed you how to generate a result set that will work just as well.
How can I do that in a single shot and get all totals for each date range. Please help me with sql.I hope this answers your question.
If not, post a little sample data (CREATE TABLE and INSERT statements) and the results you want from that data, as Saubhik requested.
Include examples of holidays: it's very unclear how you want to handle them. What is a "holday period"? How do you know if a given date is in one or not?
Simplify the problem as much as possible. For example, post a problem that gets the last 3 weeks, instead of the last 53 weeks. Anyone who suggests a solution will show you how to change it for 53 weeks, if necessary.