Update with except conditon

  • drew.allen (1/4/2012)


    ChrisM@home (1/4/2012)


    say there are ten rows with the same VND_ID and only one of them matches the filter?

    You'll need to post sample data illustrating the problem and the desired results.

    Drew

    Like this:

    DROP TABLE #PLU_1

    CREATE TABLE #PLU_1 (PLU_1ID INT IDENTITY (1,1), VND_ID INT, DSPL_DESCR VARCHAR(40))

    INSERT INTO #PLU_1 (VND_ID, DSPL_DESCR)

    SELECT 1, 'quick brown fox jumped over the lazy dog' UNION ALL

    SELECT 1, 'quick brown fox jumped over the lazy dog' UNION ALL

    SELECT 1, 'quick brown fox jumped over the lazy dog' UNION ALL

    SELECT 1, 'quick brown fox jumped over the lazy do"' UNION ALL

    SELECT 2, 'quick brown fox jumped over the lazy dog' UNION ALL

    SELECT 2, 'quick brown fox jumped over the lazy dog' UNION ALL

    SELECT 3, 'quick brown fox jumped over the lazy "' UNION ALL

    SELECT 4, 'quick brown fox jumped over the lazy do"'

    -- excludes all rows with VND_ID = 1

    -- because 1 row found in subselect

    SELECT *, replace ([DSPL_DESCR],'"','in')

    FROM #PLU_1

    where VND_ID NOT IN (

    select VND_ID from #PLU_1

    Where len(replace ([DSPL_DESCR],'"','in'))>=40

    AND DSPL_DESCR LIKE '%"%'

    )

    It's the original query, but as you figured out, it doesn't fit the requirement.


    [font="Arial"]Low-hanging fruit picker and defender of the moggies[/font]

    For better assistance in answering your questions, please read this[/url].


    Understanding and using APPLY, (I)[/url] and (II)[/url] Paul White[/url]

    Hidden RBAR: Triangular Joins[/url] / The "Numbers" or "Tally" Table: What it is and how it replaces a loop[/url] Jeff Moden[/url]

  • ChrisM@home (1/5/2012)


    drew.allen (1/4/2012)


    ChrisM@home (1/4/2012)


    say there are ten rows with the same VND_ID and only one of them matches the filter?

    You'll need to post sample data illustrating the problem and the desired results.

    Drew

    Like this:

    The problem with that data is that it violates database normalization rules, which is entirely possible, but I would only code for that possibility if I could confirm that the database does indeed violate normalization (and curse the original designer if that were the case).

    The OP already stated that the version without the subquery solves his problem confirming my assumption that the database was normalized correctly.

    Drew

    J. Drew Allen
    Business Intelligence Analyst
    Philadelphia, PA

  • drew.allen (1/5/2012)


    ChrisM@home (1/5/2012)


    drew.allen (1/4/2012)


    ChrisM@home (1/4/2012)


    say there are ten rows with the same VND_ID and only one of them matches the filter?

    You'll need to post sample data illustrating the problem and the desired results.

    Drew

    Like this:

    The problem with that data is that it violates database normalization rules, which is entirely possible, but I would only code for that possibility if I could confirm that the database does indeed violate normalization (and curse the original designer if that were the case).

    The OP already stated that the version without the subquery solves his problem confirming my assumption that the database was normalized correctly.

    Drew

    Actually, I used VND_ID subquery to isolate(not use) the VND_ID whose lenght will increase to more than 40 in the result set...

    But then I acheived the same thing by

    WHERE LEN(REPLACE(PLU.[DSPL_DESCR], '"', 'in') ) <= 40

    Regards,

    Skybvi

    Regards
    Sushant Kumar
    MCTS,MCP

  • drew.allen (1/5/2012)


    ChrisM@home (1/5/2012)


    drew.allen (1/4/2012)


    ChrisM@home (1/4/2012)


    say there are ten rows with the same VND_ID and only one of them matches the filter?

    You'll need to post sample data illustrating the problem and the desired results.

    Drew

    Like this:

    The problem with that data is that it violates database normalization rules, which is entirely possible, but I would only code for that possibility if I could confirm that the database does indeed violate normalization (and curse the original designer if that were the case).

    The OP already stated that the version without the subquery solves his problem confirming my assumption that the database was normalized correctly.

    Drew

    You've lost me Drew - I understand that the query I posted is incorrect, but how does the (theoretical) sample data I posted violate normalization? Surely it's just a foreign key?


    [font="Arial"]Low-hanging fruit picker and defender of the moggies[/font]

    For better assistance in answering your questions, please read this[/url].


    Understanding and using APPLY, (I)[/url] and (II)[/url] Paul White[/url]

    Hidden RBAR: Triangular Joins[/url] / The "Numbers" or "Tally" Table: What it is and how it replaces a loop[/url] Jeff Moden[/url]

  • ChrisM@home (1/5/2012)


    You've lost me Drew - I understand that the query I posted is incorrect, but how does the (theoretical) sample data I posted violate normalization?

    I didn't notice that you had set up an identity column.

    Surely it's just a foreign key?

    Is it? The OP never posted DDL, so it could just as easily be a primary key rather than a foreign key. It's not clear, which is why I asked whether it could be done without the subquery rather than stating outright that it could be done without the subquery.

    In fact, since the main query and the subquery use both the same table and the same column, it's more likely to be the primary key than a foreign key.

    Drew

    J. Drew Allen
    Business Intelligence Analyst
    Philadelphia, PA

  • This thread illustrates so very well the consequences of not forcing the OP to provide a proper set of requirements before answering ... talented people who could tackle the problem in just one or two posts, given accurate information, are instead left debating what the requirements actually are, many posts later, and wasting their valuable time.

    Rant over.

    The absence of evidence is not evidence of absence.
    Martin Rees

    You can lead a horse to water, but a pencil must be lead.
    Stan Laurel

Viewing 6 posts - 16 through 20 (of 20 total)

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