update only one field in a table with concatenation

  • Hi I have two tables.

    The first table has the following fields:

    (pk)id, name, ticket_details

    The second table has these two fields:

    (fk)id, ticket_details, update_time

    In the second table there can be multiple updates

    I'd like to produce a query that gives me a one line item of each record in the first table. I would also like to see all additional ticket details that have been added in the second table that is bound by the id field.

    So I want to see either in a view or a temp table:

    id, name ticket_details (from tbl1) + ticket_details(from tbl2) + update_time

    This is what I've got so far

    with cte as

    (

    select *,

    row_number() over(partition by b.id order by b.update_time desc) as rn

    tbl1 as a

    left join tbl2 as b

    on a.id = b.id

    )

    select *

    into #tcktdetails

    from cte

    declare @rn as int

    set @rn = 4

    select @rn = max(rn) from #tcktdetails where id = id

    --show last 4 updates only

    while @rn <= 4

    begin

    update a

    set a.ticket_details = a.pldetails + ' !! ' + convert(varchar, b.update_time, 114) ' !! ' + b.pldetails

    from #tcktdetails as a

    left join #tcktdetails as b

    on a.id = b.id

    where b.rn >= @rn

    set @rn = @rn +1

    end

    select *

    from #tcktdetails

    The problem is that I am only getting the last update from tbl2 to update and it adds it multiple times depending on the quantity of the rn. I'd like to add the updates only once per update.

    I hope this makes sense,

    Thanks in advance for any help with this query or a better way to do it and any other suggestions.

    Marcus.

  • Have you tried to join them?

    SELECT *

    FROM TABLE1 a

    INNER JOIN TABLE2 b ON A.id = B.id

  • Hi Adam, thanks for the reply. Yes I can join them but then I'll get a line record for every update, so if there are four updates on a record I'll end up with 4 line records. What I'd like to do is take all the update statements in tbl2 and add them to the ticket_details field in tbl1 in a query to produce a report that shows one line per record but includes all ticket details in the ticket details fields from both tbl1 + tbl2.

  • After reading the post again, my previous post did not give you the result you are trying to achieve.

    Let me look at it a little harder.

  • I just read you post again :hehe:

    Let me have another look.

    -Adam

  • Thanks Adam, I appreciate your time. I know what I'm asking for is convoluted. It's the format my boss wants the report, but I'm having trouble doing what he wants.

  • Adam, I'm trying to make it a little less complicated, is there a way to take the first update from tbl2 and then join tbl2 to itself to update the first update, with all subsequent udpates.

    So here is a sample record set.

    id details uptime rn

    a up1 time1 1

    a up2 time2 2

    a up3 time3 3

    a up4 time4 4

    The end result I would like to see in a temp table or view would be:

    id details

    a up1 time1, up2 time2, up3 time3, up4 time4

    I hope this makes what I was originally trying to do and explain easier.

  • This sample file should help you get to where you need to go. I have successfully tested it on my machine.

    declare @tbl1 table(

    id int,

    descr varchar(25)

    )

    declare @tbl2 table(

    id int,

    descr varchar(25)

    )

    insert into @tbl1

    select 1, 'test1' union all

    select 2, 'test2'

    insert into @tbl2

    select 1, 'test1 additional' union all

    select 1, 'test1 additional2' union all

    select 2, 'test2 additional'

    DECLARE RunningTotalCursor

    CURSOR LOCAL FAST_FORWARD FOR

    SELECT id, descr

    FROM @tbl2

    --ADD WHERE AND ORDER BY CLAUSE HERE IF NEEDED

    ORDER BY id

    OPEN RunningTotalCursor

    --I am not sure of the datatypes you need, so change accordingly

    DECLARE @id int

    DECLARE @descr varchar(25)

    --set starting value of the running total

    DECLARE @prevID INT

    SET @prevID = -1

    DECLARE @STR varchar(500)

    SET @STR = ''

    --hold our results data

    DECLARE @Results TABLE

    (

    id varchar(20),

    descr varchar(500)

    )

    FETCH NEXT FROM RunningTotalCursor

    INTO @id, @descr

    WHILE @@FETCH_STATUS = 0

    BEGIN

    -- if it is the first u_subconta dont worry about checking this

    -- otherwise compare the previous subconta with the current to see if

    -- we need to start from scratch.

    IF @id <> @prevID

    BEGIN

    SET @STR = ''

    END

    SET @STR = @STR + @descr +',' + ' '

    SET @prevID = @ID

    --insert values into the results table

    INSERT @Results

    VALUES (@id, @STR)

    FETCH NEXT FROM RunningTotalCursor

    INTO @id, @descr

    END

    CLOSE RunningTotalCursor

    DEALLOCATE RunningTotalCursor

    select a.id, a.descr,b.descr

    from @tbl1 a inner join

    (select id, max(descr) [descr]

    from @results

    group by id

    ) AS b on a.id = b.id

  • Please note that this particular solution uses a cursor, against my better judgement :w00t:

    If you have a substantial amount of records this query could perform very poorly, as it goes through the result set one record at a time.

  • Thanks Adam,

    Although I've read up on them a bit, I've never used a cursor before. I'll give it a try and let you know how it goes.

  • It behaves just like a cte which also walk through a set of data one row at a time.

  • Forgot to mention that some of the notation is irrelevant. This is a cursor that I helped someone else on this forum with and forgot to delete my comments.

    just an FYI. 😉

  • You can also use what is calledl the poor man's cursor to loop through the recordset as well. It's a little bit more code but is almost as efficient and less resources.

    DECLARE @ASSESSMENT_VARIABLES TABLE (

    RowNum INT IDENTITY,

    ASSESSMENT_NAMEVARCHAR(25),

    SUBJECT_CDVARCHAR(5),

    SUBJECT_DESCVARCHAR(50)

    )

    INSERT INTO @ASSESSMENT_VARIABLES

    SELECT ASSESSMENT_NAME_SHORT, SUBJECT_CD, SUBJECT_DESC

    FROM WK_ASSESSMENT_SUBJECT, WK_ASSESSMENT_TYPES

    SELECT @RowCount = 1

    SELECT @MaxRows = COUNT(*) FROM @ASSESSMENT_VARIABLES

    WHILE @RowCount <= @MaxRows

    BEGIN

    SELECT

    @SUBJECT_CD = SUBJECT_CD,

    @SUBJECT_DESC = SUBJECT_DESC

    FROM @ASSESSMENT_VARIABLES

    WHERE RowNum = @RowCount

    Code to execute

    SELECT @RowCount = @RowCount + 1

    END

    Happy coding

    Marvin

    Marvin Dillard
    Senior Consultant
    Claraview Inc

  • Since you're using 2005, you might try this method (I stole the table variable code from Adam Haines, since I'm lazy), tweaking it as needed:

    declare @tbl1 table(id int,descr varchar(25))

    declare @tbl2 table(id int,descr varchar(25))

    insert into @tbl1

    select 1, 'test1' union all

    select 2, 'test2'

    insert into @tbl2

    select 1, 'test1 additional' union all

    select 1, 'test1 additional2' union all

    select 2, 'test2 additional'

    SELECT

    t1.id

    ,Stuff((',' + t1.descr + (SELECT ',' + descr FROM @tbl2 AS t2 WHERE t1.id = t2.id FOR XML PATH(''))),1,1,'')

    FROM

    @tbl1 AS t1

    GROUP BY

    t1.id

    ,t1.descr

  • Cool solution! Using xquery to parse the string is great. Good job, this code should be substantially faster.

Viewing 15 posts - 1 through 15 (of 30 total)

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