I agree with Lieven's answer create a table containing all the months you could ever require, and use that to "LEFT JOIN" to your results table. Remember, this is a really tiny table, only 365(ish) rows per year of data you have... And you can easily write some code to populate this table initially
We do this here, and it gives lots of benefits, for example, imagine a monthly data table with the following fields (and any others you can think of!) fully populated for all the months in a given range;
- Date (E.g. 2009-04-01)
- Day (E.g. 1)
- Day of Week (E.g. Wednesday)
- Month (E.g. 4)
- Year (E.g. 2009)
- Financial Year (E.g. 2009/10)
- Financial Quarter (E.g. 2009Q1)
- Calendar Quarter (E.g. 2009Q2)
Then combining this with your query above, as follows;
SELECT `DT`.`myYear`, `DT`.`myMonth`,
AVG(`myTable`.`value1`) as avg_value_1,
AVG(`myTable`.`value2`) as avg_value_2
FROM `dateTable` as DT
LEFT JOIN `myTable`
ON `dateTable`.`myDate` = `myTable`.`save_date`
WHERE `dateTable`.`myDate` BETWEEN '2009-01-01' AND '2009-07-01'
GROUP BY `DT`.`myYear`, `DT`.`myMonth`
There may be some errors in my SQL code as I haven't been able to create the test tables, but hopefully you'll get the principal and alter to suit your needs!
Using this, you can change your "GROUP BY" clause to whatever you have in the "dateTable" table, which can allow you to easily report by Financial Quarter, Month, Day, Day of Week, etc.
Hope that helps!
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…