SQL query with pivot gives error - sql

Question

I have the following SQL query:
SELECT *
FROM (
SELECT ClientNumber, ClientName, YearEndDate, schedule, a1, ad1, a2, ad2, a3, ad3,
cv.[id] as CvId,
[dataCV] =
CASE cv.[id]
WHEN 'DATUM_AVA' THEN CONVERT(DATETIME, cv.[data], 112)
WHEN 'RECHTSVORM' THEN cv.[data]
END,
DATEADD(day, 30, [data]) AS Stat_Filing_Date,
To_Be_Filed =
CASE
WHEN DATEADD(day, 30, [data]) < GETDATE() THEN 'Yes'
ELSE 'No'
END
FROM EngagementFiles ef
INNER JOIN ac
ON ef.id = ac.EngagementFile_Id
INNER JOIN cv
ON ef.id = cv.EngagementFile_Id
RIGHT JOIN sh
ON ac.short_name = sh.a3 OR ac.short_name = sh.a2 OR ac.short_name = sh.a1
WHERE (schedule = 'G21-9A' OR schedule = 'G21-9B') AND sh.EngagementFile_Id = ef.id AND (ac.act_type = 'C' OR ac.act_type = 'H')
AND schedule =
CASE WHEN
(
SELECT cv.[data]
FROM cv
WHERE [id] = 'REPORT_TYPE' AND cv.EngagementFile_Id = ef.id
) LIKE '%VKT%' THEN 'G21-9B' ELSE 'G21-9A' END
AND ((cv.[id] = 'DATUM_AVA' AND cv.[form] = '' AND cv.[group] = ''))
) AS s
PIVOT (
MAX(dataCV)
FOR [CvId] IN ([DATUM_AVA]
)
)AS pvt
This returns the following result:
Now I want to add another column in the pivot part which also comes from the cv table. The structure of this table is kind of weird. It has the following columns:id, form, group and data.
The columns id, form and group are actually some kind of unique id which is used in a different application.
The column i want to add contains the data from the cv table where id = RECHTSVORM and group and form are empty.
The query with which I tried this is the following:
SELECT *
FROM (
SELECT ClientNumber, ClientName, YearEndDate, schedule, a1, ad1, a2, ad2, a3, ad3,
cv.[id] as CvId,
[dataCV] =
CASE cv.[id]
WHEN 'DATUM_AVA' THEN CONVERT(DATETIME, cv.[data], 112)
WHEN 'RECHTSVORM' THEN cv.[data]
END,
DATEADD(day, 30, [data]) AS Stat_Filing_Date,
To_Be_Filed =
CASE
WHEN DATEADD(day, 30, [data]) < GETDATE() THEN 'Yes'
ELSE 'No'
END
FROM EngagementFiles ef
INNER JOIN ac
ON ef.id = ac.EngagementFile_Id
INNER JOIN cv
ON ef.id = cv.EngagementFile_Id
RIGHT JOIN sh
ON ac.short_name = sh.a3 OR ac.short_name = sh.a2 OR ac.short_name = sh.a1
WHERE (schedule = 'G21-9A' OR schedule = 'G21-9B') AND sh.EngagementFile_Id = ef.id AND (ac.act_type = 'C' OR ac.act_type = 'H')
AND schedule =
CASE WHEN
(
SELECT cv.[data]
FROM cv
WHERE [id] = 'REPORT_TYPE' AND cv.EngagementFile_Id = ef.id
) LIKE '%VKT%' THEN 'G21-9B' ELSE 'G21-9A' END
AND ((cv.[id] = 'DATUM_AVA' AND cv.[form] = '' AND cv.[group] = '')
OR (cv.[id] = 'RECHTSVORM' AND cv.[form] = '' AND cv.[group] = ''))
) AS s
PIVOT (
MAX(dataCV)
FOR [CvId] IN ([DATUM_AVA],
[RECHTSVORM]
)
)AS pvt
But this gives the following error:
Conversion failed when converting date and/or time from character
string.
I'm guessing this is because for the RECHTSVORM column it is also trying to convert cv.data to a datetime.
What do I need to change in order for the query to only convert cv.data to datetime for the DATUM_AVA column?

Answer 1

It's possible that one of the Data values in the 'DATUM_AVA' rows is not in the correct format for a 112 style DATETIME. In this case 'yyyymmdd'. See CAST and CONVERT in the MSDN. You can check by selecting from the cv table with the same predicates and checking the values.
Irrespective of the above the column you are creating, dataCV, cannot be more than one data type. You are potentially trying to insert a DATETIME value and values with the originating column data type in to the created dataCV column. If you need to know that it is a DATETIME value, add an additional column to the pivot to help you identify the potential data type. You can then convert afterwards.