Question

How do I remove duplicates from a table that is set up in the following way?

unique_ID | worker_ID | date | type_ID

A worker can have multiple type_ID’s associated with them and I want to remove any duplicate types. If there is a duplicate, I want to remove the type with the most recent entry.

Answer 1

A textbook candidate for the window function row_number():

;WITH x AS (
    SELECT unique_ID
          ,row_number() OVER (PARTITION BY worker_ID,type_ID ORDER BY date) AS rn
    FROM   tbl
    )
DELETE FROM tbl
FROM   x
WHERE  tbl.unique_ID = x.unique_ID
AND    x.rn > 1

This also takes care of the situation where a set of dupes on (worker_ID,type_ID) shares the same date.
See the simplified demo on data.SE.

Update with simpler version

Turns out, this can be simplified: In SQL Server you can delete from the CTE directly:

;WITH x AS (
    SELECT unique_ID
          ,row_number() OVER (PARTITION BY worker_ID,type_ID ORDER BY date) AS rn
    FROM   tbl
    )
DELETE x
WHERE  rn > 1