Count for multiple value in table without union all in sql - sql

Question

I have validationErrors with various WIPReason associated with and I want to count based on different WIPReason, I am using Union All and it works but Query is very big with mutiple wip reason more than 10. Please find the query below
SELECT
COUNT(tlv.TransactionLineId) AS TotalErrors,
COUNT(tl.Id) AS TotalLines,
COUNT(tlv.Reason) AS NoWorkRecords,
0 AS ValidationErrors
FROM
dbo.TimesheetCellTransactionLine tctl
INNER JOIN
dbo.TransactionLine tl ON tctl.TransactionLineId = tl.Id
INNER JOIN
dbo.TransactionLineValidation tlv ON tl.Id = tlv.TransactionLineId
INNER JOIN
dbo.WIPReason w ON tlv.Reason = w.Id
WHERE
tl.CurrentStatus = 1
AND w.Id = 4 -- NoWorkRecords
GROUP BY
tlv.TransactionLineId, tl.Id
UNION ALL
SELECT
COUNT(tlv.TransactionLineId) AS TotalErrors,
COUNT(tl.Id) AS TotalLines,
0 AS NoWorkRecords,
COUNT(tlv.Reason) AS ValidationErrors
FROM
dbo.TimesheetCellTransactionLine tctl
INNER JOIN
dbo.TransactionLine tl ON tctl.TransactionLineId = tl.Id
INNER JOIN
dbo.TransactionLineValidation tlv ON tl.Id = tlv.TransactionLineId
INNER JOIN
dbo.WIPReason w ON tlv.Reason = w.Id
WHERE
tl.CurrentStatus = 1
AND w.Id = 1 -- validationErrors
GROUP BY
tlv.TransactionLineId, tl.Id
Is there any other elegant way of doing this for w.Id = 1 to 10
AND w.Id = 1 -- validationErrors
Update :
I want result as columns 10 count columns as I am using this in another big select.

Answer 1

You can use a case statement inside count like this:
SELECT
COUNT(tlv.TransactionLineId) AS TotalErrors,
COUNT(tl.Id) AS TotalLines,
COUNT(case when w.Id = 4 then tlv.Reason else null end) AS NoWorkRecords,
COUNT(case when w.Id = 1 then tlv.Reason else null end) AS ValidationErrors,
... Repeat for remaining w.Id's ...
FROM
dbo.TimesheetCellTransactionLine tctl
INNER JOIN
dbo.TransactionLine tl ON tctl.TransactionLineId = tl.Id
INNER JOIN
dbo.TransactionLineValidation tlv ON tl.Id = tlv.TransactionLineId
INNER JOIN
dbo.WIPReason w ON tlv.Reason = w.Id
WHERE
tl.CurrentStatus = 1
GROUP BY
tlv.TransactionLineId, tl.Id
Note
Since you are using an inner join on tl.Id = tlv.TransactionLineId
TotalErrors/TotalLines are always going to be the same
Grouping by tl.Id and tlv.TransactionLineId is unnecessary

Answer 2

SELECT COUNT(tlv.TransactionLineId) OVER (partition by (w.Id) order by w.Id) AS TotalErrors,
COUNT(tl.Id) OVER (partition by (w.Id) order by w.Id) AS TotalLines,
0 AS NoWorkRecords,
COUNT(tlv.Reason) OVER (partition by (w.Id) order by w.Id) AS ValidationErrors
FROM dbo.TimesheetCellTransactionLine tctl
INNER JOIN dbo.TransactionLine tl ON tctl.TransactionLineId = tl.Id
INNER JOIN dbo.TransactionLineValidation tlv ON tl.Id = tlv.TransactionLineId
INNER JOIN dbo.WIPReason w ON tlv.Reason = w.Id
WHERE tl.CurrentStatus = 1
AND w.Id between 1 and 10 -- validationErrors
GROUP BY tlv.TransactionLineId,
tl.Id,tlv.Reason