SQL: Replacing NULL with 0 in monthly count table

Question

I have a table named GAINLP The table contains fields

'Record#"           (INT), 
'SightingDate'      (DATE), 
'SpeciesName'       (VARCHAR)

Need SQL to output an array that contains an integer that corresponds to the sum of SightingDate for each month.

Example : 0,0,0,0,1,5,10,12,5,3,0,0

Instead, the following code causes null value sums to be skipped and I'm left with 1,5,10,12,5,3

`select count(SightingDate) from GAINLP where SpeciesName LIKE '%Actias luna' GROUP BY MONTH(SightingDate)`

I understand that this can be done by joining with a calendar table, but I've not found examples of such code that also employs the WHERE operator.

Answer 1

You seem to be looking for conditional aggregation. The principle is to move the filtering logic to a CASE construct within the aggregate function.

SELECT 
    SUM(
        CASE
            WHEN SpeciesName LIKE '%Actias luna'
                THEN 1 
            ELSE 0
        END
    )
GROM gainlp 
GROUP BY MONTH(SightingDate)

Answer 2

You could use conditional aggregation here:

SELECT
    MONTH(SightingDate) AS month,
    COUNT(CASE WHEN SpeciesName LIKE '%Actias luna' THEN 1 END) cnt
FROM GAINLP
GROUP BY
    MONTH(SightingDate);

But, this might not be as efficient as a calendar table based join approach. Here is how you might do that:

SELECT
    t1.month,
    COUNT(t2.SightingDate) cnt
FROM (SELECT DISTINCT MONTH(SightingDate) AS month FROM GAINLP) t1
LEFT JOIN GAINLP t2
    ON t1.month = MONTH(t2.SightingDate)
WHERE
    t2.SpeciesName LIKE '%Actias luna'
GROUP BY
    t1.month;

Note: It might be possible that your data could span more than a given year, in which case you would probably want to report both the year and month.