Question

I have a table that has duplicate email address, I need to insert just one of them into a temp a temp table along with two other fields. there are many example here but I can get any of them work,

I ended up looking into MERGE I get the same results. All the record are getting inserted I’m at a lost. I tried many different samples but it always insert all the records. I went back to make sure the email address are really dupes and they are.. Below is were I’m at now..

MERGE #EmailTable2 AS Target
USING (SELECT EMAIL, NAME, JOB_TITLE FROM b2b_cmas_list$ WHERE EMAIL IS NOT NULL) AS Source
ON (Target.EMAIL = Source.EMAIL)
 WHEN NOT MATCHED BY TARGET THEN
    INSERT (EMAIL, NAME, JOB_TITLE)
    VALUES (Source.EMAIL, Source.NAME, Source.JOB_TITLE)

OUTPUT $action, inserted.*, deleted.*;

so any help in getting this correct would be helpful.

Answer 1

This it not working because SQL doesn’t know, which of the two rows containing the same e-mail you want to choose. I mean: if EMAIL is the same, which of pair (NAME and JOB_TITLE) are important and which can be discarded?

Some hints:

1. If it doesn’t matter which item is chosen simply group by EMAIL selecting MAX(NAME) and MAX(JOB_TITLE), i.e.

   select EMAIL, max(NAME), max(JOB_TITLE) from b2b_cmas_list$ group by EMAIL

   Be warned however that this can mangle NAME-JOB_TITLE pairs (as max is selected).

2. Try using ROW_NUMBER() OVER() to arbitrarilly select 1st row in each group.

3. Use a CURSOR to iterate over rows and skip duplicates.

4. Use .NET CLR aggregate to i.e. concat names and job titles for same e-mail.

And a little note to your MERGE statement. This is not working as expected, because SQL checks all rows at once, and not row-by-row. So it is not that if one e-mail. ie. “[e-mail]” is inserted, then another won’t. It only matters if “[e-mail]” is in the table at the beginning of the statement.