Update Trigger Assistance - Update Same Row More than Once?

Question

Good morning,
I'm new to Triggers and am not understanding why this is not working.  I've tried the web and none of the articles seem to answer my question.  What I'm trying to do is clean a string and standardize some data the users input into a 1000 character
varchar field.  I have a source table which has a trigger on it that inserts a limited number of columns that we'll use for reporting purposes.  After the row has been inserted into the table (referred to as "destination" in the code block),
I'm looking to update the comments string, capitalizing the first character of the string, standardizing spacing after a colon, removing hyphens at the beginning of the string.  The source and destination tables both have the comments field as a varchar
1000, not null data type with a key id between the source and destination tables.  At another segment of our system, there were errors reporting nulls and when this was removed, everything cleared up.  I can do a series of replace statements when
the data is required in reports but would rather have it clean in the destination table.  The code is below:
CREATE TRIGGER [dbo].[destination_Update_trg]
ON [dbo].[destination]
AFTER UPDATE
AS
BEGIN
SET NOCOUNT ON;
IF EXISTS (SELECT *
FROM destination C
JOIN INSERTED I
ON I.PrimaryKeyID = C.PrimaryKey
WHERE I.Comments IS NOT NULL AND I.Comments <> '')
BEGIN
UPDATE T
SET T.Comments =
CASE
WHEN T.Comments <> '' AND LEFT(CAST(T.Comments AS VARCHAR (1000)), 2) LIKE '- %'
THEN SUBSTRING(T.Comments, 3, 1000)
END
FROM destination T
JOIN INSERTED I
ON T.primarykey = I.primarykey
WHERE T.primarykey = I.primarykey
UPDATE T
SET T.Comments = UPPER(LEFT(CAST(T.Comments AS VARCHAR (1000)), 1)) + SUBSTRING(CAST(T.Comments AS VARCHAR (1000)), 2, 1000)
FROM destination T
JOIN INSERTED I
ON T.primarykey = I.primarykey
WHERE T.primarykey = I.primarykey
UPDATE T
SET T.Comments = REPLACE(SUBSTRING(T.Comments, 1, 1000), ': ', ': ')-- '[A-9]: [A-9]'
FROM destination T
JOIN INSERTED I
ON T.primarykey = I.primarykey
WHERE T.primarykey = I.primarykey
AND SUBSTRING(T.Comments, 1, 1000) LIKE '%[A-Z]%: [A-Z]%'
END
RETURN
END;
GO
Perhaps there's a better way to accomplish the task. I appreciate your review and assistance. Thank you.

Answer 1

First, the trigger only fires when you do an UPDATE.  I'm not sure from your description whether or not you are expecting this trigger to fire when you do an insert on dbo.destination.  If you want it to fire on INSERTs, then you need to specify
FOR INSERT (or FOR INSERT, UPDATE if you want it to fire on both inserts and updates).
More importantly, the first update in your trigger does
SET T.Comments =
CASE
WHEN T.Comments <> '' AND LEFT(CAST(T.Comments AS VARCHAR (1000)), 2) LIKE '- %' THEN SUBSTRING(T.Comments, 3, 1000)
END
When you do a CASE and you don't have an ELSE clause, there is an implicit ELSE NULL clause added.  So your update is equivalent to
SET T.Comments =
CASE
WHEN T.Comments <> '' AND LEFT(CAST(T.Comments AS VARCHAR (1000)), 2) LIKE '- %' THEN SUBSTRING(T.Comments, 3, 1000)
ELSE NULL
END
So your very first update sets the Comments column to NULL in every row that was updated where the Comments column is not like '- %'.  You want to change the case statement to be
SET T.Comments =
CASE
WHEN T.Comments <> '' AND LEFT(CAST(T.Comments AS VARCHAR (1000)), 2) LIKE '- %' THEN SUBSTRING(T.Comments, 3, 1000)
ELSE T.Comments
END
(Or you can put the test in the where clause so that you only update those rows where the Comments column is like '- %'.  That would look like
UPDATE T
SET T.Comments = SUBSTRING(T.Comments, 3, 1000)
END
FROM destination T
JOIN INSERTED I
ON T.primarykey = I.primarykey
WHERE LEFT(CAST(T.Comments AS VARCHAR (1000)), 2) LIKE '- %'
(Note that you do not need to include the T.Comments <> '' because if T.Comments is Like '- %', then it cannot be the empty string and you do not need to include WHERE T.primarykey = I.primarykey because you know that is true since you did an inner
join on T.primarykey = I.primarykey.
Tom