Question

I have a column called ‘comments’, and would like to find all records containing strings in the ‘comments’ column which are an exact substring of any other cell in the same column. These exact substrings must be entire words, ie they should be seperated by a space from the rest of the string of which they are a substring.

I think I need to use something similar to an array function. How is this possible in SQL Server 2005?

Answer 1

Perhaps something like this:

WITH data (id, str) AS (
  SELECT 1, 'some' UNION ALL
  SELECT 2, 'thing' UNION ALL
  SELECT 3, 'somethings' UNION ALL
  SELECT 4, 'something else' UNION ALL
  SELECT 5, 'two things' UNION ALL
  SELECT 6, 'twenty two things'
)
SELECT str
FROM data d
WHERE EXISTS (
  SELECT *
  FROM data
  WHERE str     LIKE '%'     + d.str +     '%'
    AND str NOT LIKE '%[^ ]' + d.str +     '%'
    AND str NOT LIKE '%'     + d.str + '[^ ]%'
    AND id <> d.id
)

This will find values (words or phrases) that are surrounded by spaces in another value of the same column or are at the beginning or end of another value of the same column.

This particular example produces just one result:

str
----------
two things

The query will hardly be very efficient, though, to say the least, but, hopefully, might inspire you or someone else to a better idea.