Simple Crosstab Procedure with Power

  • cgruel

    Grasshopper

    Points: 14

    Comments posted to this topic are about the item Simple Crosstab Procedure with Power

  • gjcolledge

    Grasshopper

    Points: 19

    I don't have experience using crosstabs, but shouldn't the results in your example be:

    TYPE RED BLUE GREEN

    A 2 0 0

    B 1 0 0

    C 0 1 0

    D 0 0 1

  • Mark D Powell

    SSCarpal Tunnel

    Points: 4453

    I can't get past the syntax errors. I wish ServerCentral would fix their script upload feature because getting the scripts to work has proven to be a royal pain.

    -- Mark D Powell --

    Finally, got it working. I had to retype the lead into several of the lines which display as little boxes in the posting and had become invisible when I copied and pasted via Word. Notepad shows the garbage. Once removed the procedure compiled and appears to work.

  • adammenkes

    SSChasing Mays

    Points: 602

    I am having a problem getting this to do what I want... First problem, I cannot seem to get the procedure to take a query as an argument in parenthesis as stated.

    To test, I just wrapped 'Table1' with '(SELECT * FROM Table1)'

    -- Incorrect syntax near ')'.

    The second problem is more complex and I will post another reply.

  • adammenkes

    SSChasing Mays

    Points: 602

    The second problem I am having is more complex. I am trying to figure out 2 things.

    1) How to get the column headings (and totals)

    2) How to have a report use dynamic columns

    Below is a sample table I created:

    CREATE TABLE [dbo].[TimeSheetTest](

    [ID] [int] IDENTITY(1,1) NOT NULL,

    [EmployeeCode] [varchar](50) NOT NULL,

    [ProjectCode] [varchar](50) NOT NULL,

    [HoursWorked] [decimal](18, 2) NOT NULL,

    [DateWorked] [datetime] NOT NULL,

    CONSTRAINT [PK_TimeSheetTest] PRIMARY KEY CLUSTERED

    (

    [ID] ASC

    ) ON [PRIMARY]

    ) ON [PRIMARY]

    Then, add the following data:

    INSERT INTO TimeSheetTest

    SELECT 'Adam', 'P1', 8, '2009-01-01'

    UNION ALL SELECT 'Adam', 'P2', 1, '2009-01-01'

    UNION ALL SELECT 'Bob', 'P1', 5, '2009-01-01'

    UNION ALL SELECT 'Bob', 'P2', 3, '2009-01-01'

    UNION ALL SELECT 'Adam', 'P1', 8, '2009-01-03'

    UNION ALL SELECT 'Adam', 'P1', 8, '2009-01-05'

    UNION ALL SELECT 'Bob', 'P1', 7, '2009-01-05'

    UNION ALL SELECT 'Bob', 'P2', 1, '2009-01-05'

    Now, using the basic syntax for getting the xtab:

    exec UberCrosstab

    'EmployeeCode, ProjectCode',

    'DateWorked',

    'TimeSheetTest',

    'HoursWorked',

    'SUM'

    I get the following results (I dropped the time 12:00AM from below for readability):

    EmployeeCode ProjectCode Jan 1 2009 Jan 3, 2009 Jan 5, 2009

    Adam P1 8.00 8.00 8.00

    Bob P1 5.00 NULL 7.00

    Adam P2 1.00 NULL NULL

    Bob P2 3.00 NULL 1.00

    This part works as expected. However, I want to display Thu 01/01, Sat 01/03, Mon 01/05

    so, I created a field as follows:

    SELECT *, LEFT(DATENAME(dw, DateWorked), 3) + ' ' + CONVERT(VARCHAR(5), DateWorked, 1) AS Header

    FROM TimeSheetTest

    Of course (since I cannot get the (query) to work), I do the following:

    SELECT *, LEFT(DATENAME(dw, DateWorked), 3) + ' ' + CONVERT(VARCHAR(5), DateWorked, 1) AS Header

    INTO #temp

    FROM TimeSheetTest

    exec UberCrosstab

    'EmployeeCode, ProjectCode',

    'Header',

    '#temp',

    'HoursWorked',

    'SUM'

    drop table #temp

    Now, the table appears with headings of what I want (Mon 01/05, Sat 01/03, Thu 01/01), but the problem is they are sorted by the text (M in Monday comes before S - Saturday, etc).

    [highlight=#ffff11]

    Q. How do I sort it the way I want?

    [/highlight]

    I am also trying to get all the dates in a date range, not just those that have data. This is not a problem, other than the data comes out sorted by Fri, Mon, Sat, Sun, Thurs, Tues, Wed.

    SET NOCOUNT ON

    DECLARE @StartDate DATETIME

    DECLARE @EndDate DATETIME

    DECLARE @ctr INT

    DECLARE @diff INT

    DECLARE @days TABLE (

    WorkDay DATETIME PRIMARY KEY,

    WorkDayName VARCHAR(50),

    Counter INT

    )

    SELECT @StartDate = '2009-01-01'

    SELECT @EndDate = '2009-01-07'

    SELECT @diff = DATEDIFF(d, @StartDate, @EndDate)

    SELECT @ctr = 0

    WHILE @ctr <= @diff BEGIN

    INSERT INTO @days (

    WorkDay,

    WorkDayName,

    Counter

    ) VALUES (

    @StartDate + @ctr,

    LEFT(DATENAME(dw, @StartDate + @ctr), 3) + '. ' + LEFT(CONVERT(VARCHAR(8), @StartDate + @ctr, 1), 5),

    @ctr + 1

    )

    SET @ctr = @ctr + 1

    END

    SELECT *

    INTO #temp

    FROM @days d

    LEFT OUTER JOIN TimeSheetTest t

    ON d.WorkDay = t.DateWorked

    exec UberCrosstab

    'EmployeeCode, ProjectCode',

    'WorkDayName',

    '#temp',

    'HoursWorked',

    'SUM'

    drop table #temp

    Part 2 of the problem is to create a report that will have dynamic fields in the crosstab, and a total afterwards. I have not tried it, but I am assuming I can create another table variable or query that totals across for the period (in this case, one week) and join that with this data. That will get me the totals.

    The tricky part is the workdays. It might be 7. It might be 14 (paid every 2 weeks). It might be a little more if paid twice a month (1-15, 16-31 for some months, 16-28 or 29 for February, etc). It could be 28-31 days (depending again on the month).

    [highlight=#ffff11]

    Please tell me there is a way other than to create 31 columns and hide the ones that are not used.

    [/highlight]

  • pgarty

    Valued Member

    Points: 59

    The problem is a subquery is being used but not aliased. It is when the variable @query is being defined

    To fix it change to:

    set @query = '

    DECLARE @cols NVARCHAR(max)

    SELECT @cols = STUFF(

    ( SELECT DISTINCT ' + case when @colTop is not null then 'Top ' + @colTop else '' end + '''],['' + t2.' + @pivotField + '

    FROM

    (SELECT CONVERT(nvarchar(max),a.' + @colField + ') AS ' + @pivotField + '

    FROM

    (SELECT DISTINCT ' + @colField + '

    FROM ' + @colTable + ' x

    ' + case when @colWhere is not null then 'WHERE ' + @colWhere else '' end + '

    ) AS a

    ) AS t2

    ORDER BY ''],['' + t2.' + @pivotField + ' FOR XML PATH('''')

    ), 1, 2, '''') + '']''

    select @cols

    '

    All I have added is ' x '

    Hope this helps

    Phil

  • pgarty

    Valued Member

    Points: 59

    PS that was for adammenkes first problem

  • akljfhnlaflkj

    SSC Guru

    Points: 76202

    Right or wrong, it looks like a chance for me to practice work with PIVOT.

Viewing 8 posts - 1 through 8 (of 8 total)

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