Question

im sure this is easy but im having a block I am trying to write some sql against a single table that has (simplified for example)

RunName, Result
foo,     pass
foo,     pass
foo,     fail 
foo,     pass
boo,     pass
boo,     fail
boo,     fail
soo,     pass

I was a query that will return a count of pass or fail for each name

something like for fail

foo, 1
boo, 2
soo, 0

and for pass

foo, 3
boo, 1
soo, 1

Answer 1

Normally you’d do a simple COUNT and GROUP BY RunName, but that will not show “zero results” like soo’s fails.

This should work even for zero results (on MySQL)

SELECT RunName, SUM(Result='fail')
FROM TableA
GROUP BY RunName;

Demo here.

Edit: As Aaron points out, that only works on MySQL, this works on SQL Server also;

SELECT  RunName, SUM(CASE WHEN Result = 'fail' THEN 1 ELSE 0 END) as fails
FROM   TableA
GROUP BY RunName

Demo here.

Edit2: As Marcus points out, it’s may not be a very index friendly way of doing the query, so if you have a primary key you may be better off doing a self join with a count/group to get the correct result;

SELECT a.RunName, COUNT(b.Result)
FROM TableA a LEFT JOIN TableA b ON a.id=b.id AND b.Result='fail'
GROUP BY RunName

Demo here.