SQL Server Function: Cursor or Tally Table; which to use, HELP

  • Hi lowell

    I've tested the code and it worked beautifully. I need a little help though, I got the logic wrong when I explained the issue originally. There should never be any instance of the pattern not starting with a 1;however, the very first row generated needs to be the very first instance of the start date, not the start date with 7 adds added. I have tried to get around this by slightly altering the case statement to replace the first row date with the actual start date but then the second row date that is generated doesn't use the replace start date from the first row but the original, meaning that it's out by a week.

    Any ideas?

    Here's the code I've changed

    ;WITH

    TENS (N) AS (SELECT 0 UNION ALL SELECT 0 UNION ALL SELECT 0 UNION ALL

    SELECT 0 UNION ALL SELECT 0 UNION ALL SELECT 0 UNION ALL

    SELECT 0 UNION ALL SELECT 0 UNION ALL SELECT 0 UNION ALL SELECT 0),

    THOUSANDS (N) AS (SELECT 1 FROM TENS t1 CROSS JOIN TENS t2 CROSS JOIN TENS t3),

    TALLY (N) AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT 0)) FROM THOUSANDS)

    ,

    TestCTE(reference,StartDate,Pattern,PatternLength) AS

    (

    SELECT 'VideoGames-S1',convert(datetime,'01/01/2013 10:45'),'11 111 111111  1111','18' --Course 13NAPUSD-B2

    )

     

    SELECT

                                    'StartTime' =

                                    CASE

                                    WHEN Tally.N = 1 THEN [startdate]

                                    WHEN SUBSTRING([pattern],N,1) = '1'  THEN DATEADD(wk,Tally.N,[startdate]) ELSE NULL 

                                    END,

                                    TestStartDate_DATEADD = DATEADD(wk,Tally.N,[startdate])

                                    ,*

                                    , SUBSTRING([pattern],N,1)

                                   

    FROM

                                    TestCTE testcte

    CROSS JOIN Tally tally

     

    WHERE

                                    N <= LEN([pattern])

    --AND SUBSTRING([pattern],N,1) = '1'

     

    ORDER BY

                                    reference, N

  • knives the fix is rediculously easy;

    the Tally table needs to start at zero instead of one.

    so this tiny subsection changes:

    TALLY (N) AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT 0)) FROM THOUSANDS)

    --change to

    TALLY (N) AS (SELECT ROW_NUMBER() OVER (ORDER BY (SELECT 0)) - 1 FROM THOUSANDS)

    Lowell


    --help us help you! If you post a question, make sure you include a CREATE TABLE... statement and INSERT INTO... statement into that table to give the volunteers here representative data. with your description of the problem, we can provide a tested, verifiable solution to your question! asking the question the right way gets you a tested answer the fastest way possible!

  • Hi lowell

    I've managed to sort it now, figured it out about 2 mins after posting. Having a blond day today.

    In the THEN clause I just -7 after the calculation. This moves all predicted dates back 7 days.

    Sorry for not updating post, only just had a spare minute.

    Thanks for the help, it's appreciated 🙂

Viewing 3 posts - 16 through 17 (of 17 total)

You must be logged in to reply to this topic. Login to reply