Word count for all the words appearing in a column in SQL Server 2008

Question

I have a table called 'ticket_diary_comment' with a column called 'comment_text'. This column is populated with text data. I would like to get the frequency of all the words occurring in this entire column. Ex:

Comment_Text
I am a good guy
I am a bad guy
I am not a guy

What I want:

Word    Frequency
I       3
good    1
bad     1
not     1
guy     3

Notice that I have also removed the stop words in the output. I know calculating the frequency of a particular word is not difficult but I am looking for something that counts all the words appearing in a column removing the stop words.

I would appreciate any kind of help on this issue. I would also like to mention that I have to apply this query on a big-ish dataset (about 1 TB), so performance is a concern.

Answer 1

I would use a table valued function to split the strings, and then group them in a query. Something like this:

SELECT item, count(1)
FROM ticket_diary_comment 
    CROSS APPLY dbo.fn_SplitString(comment_text, ' ')
GROUP BY item

and the definition for fn_SplitString:

CREATE FUNCTION [dbo].[fn_SplitString]   
(   
    @String VARCHAR(8000),   
    @Delimiter VARCHAR(255)   
)   
RETURNS   
@Results TABLE   
(   
    ID INT IDENTITY(1, 1),   
    Item VARCHAR(8000)   
)   
AS   
BEGIN   
INSERT INTO @Results (Item)   
SELECT SUBSTRING(@String+@Delimiter, num,   
    CHARINDEX(@Delimiter, @String+@Delimiter, num) - num)   
FROM Numbers   
WHERE num <= LEN(REPLACE(@String,' ','|'))   
AND SUBSTRING(@Delimiter + @String,   
            num,   
            LEN(REPLACE(@delimiter,' ','|'))) = @Delimiter   
ORDER BY num RETURN   
END   

This function requires a numbers table, which is basically just CREATE TABLE Numbers(Num int) and contains all the numbers from 1 to 10,000 (or more/less depending on needs). If you already have a numbers table in your DB you can substitute that table/column for what you already have.