I’m struggling with a query where I need to join data.
I have a table with all kinds of status messages. These have a failureBitNr like for example “a1_1”. Then I have a second table that translates “a1_1” to a human readable text (lng_English). By default, the second table has a translation for “a1_1”. But this can also be overwritten by another specific translation, for example when the failureBitNr is used for another status than standard.
So I need to join the status table with the translation table on failureBitNr. That’s not that hard.
But how do I join it on the translation that overwrites the basic translation?
Table 2 looks like this:
id fileProjectNr MachineNr failureBitNr lng_English
905 2203456 2 a6_1 This message overwrites the standard
205 NULL NULL a6_1 Standard message for a6_1
204 NULL NULL a1_1 Standard message for A1_1
203 NULL NULL a1_2 Standard message for A1_2
Remark that the message that overwrites the existing one holds a fileProjectNr that is not NULL. All standard messages have a fileProjectNr NULL
So joining only on FailureBitNr will return both rows (905 & 205). But I need to join on failureBitNr and do something with not being NULL of fileProjectNr.
So I did this:
DECLARE @ProjectNr int = 123456
SELECT
t1.*,
t2.*
FROM
Table1 AS t1
LEFT JOIN
Table2 AS t2
ON
(t1.failureBitNr = t2.failureBitNr)
AND
(t2.fileProjectNr LIKE
CASE WHEN t2.fileProjectNr = @ProjectNr THEN
@ProjectNr
ELSE
NULL
END
)
WHERE
{where statement}
This will return ID 905, however if there is also a join on failureBitNr a1_1 and a1_2, these will all be returned as “NULL” in stead of “Standard message for ax_x”.
Anyone an idea of how to fix this?
My first idea: