Dynamic PIVOT CLR

  • Eric Wahner-345205

    Mr or Mrs. 500

    Points: 544

    Comments posted to this topic are about the item Dynamic PIVOT CLR

  • rmcomics

    Old Hand

    Points: 332

    Very nice article.

    A question:

    Is it possible for me to for example create a Class, generate a List of that class, populate it with the SQL Server query information, after that use LINQ to manipulate the information in this List and after return the content of this list on the procedure?

    Thanks a lot.

  • rgp151

    SSC Veteran

    Points: 297

    It's good, but it still doesn't resolve the biggest problem with SQL's PIVOT functionality (which is no fault of yours), which is that there is no way to create a dynamic pivot result-set that you can join to.

    For example, if you could create a dynamic pivot in a table valued function then you could join to the function, but there isn't a way to create a dynamic pivot UDF.

    PIVOT would be a heck of a lot more powerful if you could do that...

  • Jeff Moden

    SSC Guru

    Points: 994934

    Good article.

    I do want to point out that dynamic pivots, especially for things like "rolling date reporting" just aren't difficult to do in SQL Server.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column.
    "If you think its expensive to hire a professional to do the job, wait until you hire an amateur."--Red Adair
    "Change is inevitable... change for the better is not."
    When you put the right degree of spin on it, the number 3|8 is also a glyph that describes the nature of a DBAs job. 😉

    Helpful Links:
    How to post code problems
    Create a Tally Function (fnTally)

  • Eric Wahner-345205

    Mr or Mrs. 500

    Points: 544

    I would agree when you are using known and predictable values like a set of dates or even dates that haven't occurred, creating a PIVOT in SQL is quite simple. This exercise was for those types of collections of data that are "dynamic" and ever changing. When you have an unpredictable set of data that you need to pivot, you really have no other choice.

  • Eric Wahner-345205

    Mr or Mrs. 500

    Points: 544

    @rmcomics

    I don't see why not as long as your list is either based upon static information or populated from the same connection. Though I imagine if the data that you wanted to integrate into the list were on the same connection, that you would somehow integrate that into the select statement that is passed into the @query parameter.

    Perhaps you are just asking in general can you use LINQ within a CLR function? The answer would also be yes.

  • jcrawf02

    SSC-Insane

    Points: 24198

    Eric Wahner-345205 (12/6/2012)


    I would agree when you are using known and predictable values like a set of dates or even dates that haven't occurred, creating a PIVOT in SQL is quite simple. This exercise was for those types of collections of data that are "dynamic" and ever changing. When you have an unpredictable set of data that you need to pivot, you really have no other choice.

    Well, no, that's not true, you just have to use dynamic SQL.

    Something like this:

    /*

    Title: ExtProps.spjc_viewExtendedProperties

    Author: Jon Crawford

    Description: pivots the extended properties that are available in the database

    Business Need: to search the db for existing objects

    What is the user going to do with this?:

    Known flaws: n/a

    Revision History:

    Date Changes

    ------ -------

    9/13/2012 'initial implementation'

    9/13/2012 modified to use dynamic SQL so that if new properties show up, they'll be included

    */

    ALTER PROCEDURE [ExtProps].[spjc_viewExtendedProperties] (@searchTerm VARCHAR(255) = NULL)

    AS

    BEGIN

    DECLARE @sql VARCHAR(MAX),

    @properties VARCHAR(8000)

    --find all the user created extended properties,

    --but to force a reasonable order into the view we use a temp table to store distinct values

    -- before you use them in the dynamic SQL below

    DECLARE @table TABLE (name VARCHAR(255),orderValue int)

    INSERT INTO @table (name,orderValue)

    SELECT DISTINCT name ,

    --here's where we pick the order of columns we want

    CASE name

    WHEN 'Title' THEN 1

    WHEN 'Author' THEN 2

    WHEN 'Description' THEN 3

    WHEN 'Known Flaws' THEN 4

    ELSE 99 --everything else will just show up after these ones, no particular order

    END AS orderValue

    FROM ExtProps.Properties

    --get rid of the Microsoft extended properties

    WHERE name NOT IN ('Caption','Long_Description',

    'microsoft_database_tools_support',

    'MS_Description',

    'MS_DiagramPane1',

    'MS_DiagramPaneCount')

    ORDER BY orderValue

    --==============================

    -- testing

    --SELECT *

    --FROM @table AS t

    --==============================

    --shove the names of all the distinct extended properties into a variable so we can use it in the dynamic SQL

    SELECT @properties= COALESCE(@properties ,'')+'['+CONVERT(VARCHAR(255),p.name)+'],'

    FROM @table AS p

    --get rid of the last comma that we added just above

    SET @properties = LEFT(@properties,LEN(@properties)-1)

    --==============================

    -- testing

    --SELECT @properties

    --==============================

    --force all of the unique property names into our PIVOT statement below,

    -- but hard-code the ltrim of the name and value (gets rid of leading blanks which cause ordering issues)

    -- and force the order by Title to make it look neat

    SET @sql =

    'SELECT '+@properties+'

    FROM ( SELECT p.class ,

    p.class_desc ,

    p.object_schema ,

    p.object_name ,

    p.column_name ,

    p.major_id ,

    p.minor_id ,

    LTRIM(CONVERT(varchar(255),p.name)) AS name ,

    LTRIM(CONVERT(varchar(1000),p.value)) AS value

    FROM ExtProps.Properties AS p

    ) AS sourceTable

    PIVOT ( MIN(value) FOR [name] IN ( '+@properties+' ) ) AS PivotTable

    WHERE pivotTable.Title IS NOT NULL

    AND (

    pivotTable.Title LIKE ''%'+COALESCE(@searchTerm,'')+'%''

    OR pivotTable.Description LIKE ''%'+COALESCE(@searchTerm,'')+'%''

    )

    ORDER BY [Title]'

    --==============================

    -- testing

    --PRINT @sql

    --==============================

    --now execute the PIVOT statement we built to return data

    EXEC(@SQL)

    END

    **Edit - I should have had a link in there, in case anyone is wondering, I'm pivoting the extended properties that I populate using Michael Coles' sp's (very handy)

    http://sqlblog.com/blogs/michael_coles/archive/2010/01/12/t-sql-tuesday-easy-extended-properties.aspx

    ---------------------------------------------------------
    How best to post your question[/url]
    How to post performance problems[/url]
    Tally Table:What it is and how it replaces a loop[/url]

    "stewsterl 80804 (10/16/2009)I guess when you stop and try to understand the solution provided you not only learn, but save yourself some headaches when you need to make any slight changes."

  • JustASQLGuy

    Old Hand

    Points: 387

    I've done this loads of time using TSQL sps and dynamic sql within, much in the same way as the CLR routine. Is there an advantage to using CLR over TSQL?

  • Eric Wahner-345205

    Mr or Mrs. 500

    Points: 544

    vopipari (12/6/2012)


    I've done this loads of time using TSQL sps and dynamic sql within, much in the same way as the CLR routine. Is there an advantage to using CLR over TSQL?

    There are somethings that SQL does quite well and then there are somethings that .NET does well. What my approach attempts to do is simplify it so that you don't have to write a custom procedure for every PIVOT that you do. I too have used dynamic sql to solve this problem as well, and while I haven't done any performance testing to see which one performs better than the other, I can say that there should be fewer trips to the database with the CLR method.

    This was more of an educational post to show how it could be done using a CLR procedure. I leave it up to the community to decide if this works for them or not.

  • Eric Wahner-345205

    Mr or Mrs. 500

    Points: 544

    @jcrawf02

    I think you confused my point with using a PIVOT vs dynamic PIVOT. When I said you have no choice, I was referring to doing a dynamic PIVOT. If you have unpredictable results you cannot write a PIVOT query without using dynamic sql to select the distinct rows that will ultimately be your columns.

  • Jeff Moden

    SSC Guru

    Points: 994934

    Eric Wahner-345205 (12/6/2012)


    vopipari (12/6/2012)


    I've done this loads of time using TSQL sps and dynamic sql within, much in the same way as the CLR routine. Is there an advantage to using CLR over TSQL?

    There are somethings that SQL does quite well and then there are somethings that .NET does well. What my approach attempts to do is simplify it so that you don't have to write a custom procedure for every PIVOT that you do. I too have used dynamic sql to solve this problem as well, and while I haven't done any performance testing to see which one performs better than the other, I can say that there should be fewer trips to the database with the CLR method.

    This was more of an educational post to show how it could be done using a CLR procedure. I leave it up to the community to decide if this works for them or not.

    Eric Wahner-345205 (12/6/2012)


    I would agree when you are using known and predictable values like a set of dates or even dates that haven't occurred, creating a PIVOT in SQL is quite simple. This exercise was for those types of collections of data that are "dynamic" and ever changing. When you have an unpredictable set of data that you need to pivot, you really have no other choice.

    Not quite true. It's "just" dynamic SQL and the same thing could be done there as what has been done in the managed code. Still, my hat's off to you for making life easier. I've always had a bad taste in my mouth about the current version of PIVOT especially when things like the ACCESS version of PIVOT works so very well.

    Shifting gears a bit, have you done any testing performance wise?

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column.
    "If you think its expensive to hire a professional to do the job, wait until you hire an amateur."--Red Adair
    "Change is inevitable... change for the better is not."
    When you put the right degree of spin on it, the number 3|8 is also a glyph that describes the nature of a DBAs job. 😉

    Helpful Links:
    How to post code problems
    Create a Tally Function (fnTally)

  • jcrawf02

    SSC-Insane

    Points: 24198

    Eric Wahner-345205 (12/6/2012)


    @jcrawf02

    I think you confused my point with using a PIVOT vs dynamic PIVOT. When I said you have no choice, I was referring to doing a dynamic PIVOT. If you have unpredictable results you cannot write a PIVOT query without using dynamic sql to select the distinct rows that will ultimately be your columns.

    I'm not sure how that's any different than what you did, you just passed the list of fields to the CLR where I didn't bother to pass them, just collected them from available fields in the system and stuffed them back into the SQL query I constructed. I could make mine do what yours does, without having to use a CLR.

    Nice to know how to do it that way too, and well-written article, but it's not the *only* option.

    ---------------------------------------------------------
    How best to post your question[/url]
    How to post performance problems[/url]
    Tally Table:What it is and how it replaces a loop[/url]

    "stewsterl 80804 (10/16/2009)I guess when you stop and try to understand the solution provided you not only learn, but save yourself some headaches when you need to make any slight changes."

  • Misha_SQL

    SSCertifiable

    Points: 5385

    Thank you for the great article! .NET world has always been a bit of a mystery to me (as I suspect for many other DB folks), I would only deal with it in SSIS scripts and similar circumstances. Your article is a great learning tool, not to mention the usefulness of being able to do dynamic PIVOTs.

  • Jeff Moden

    SSC Guru

    Points: 994934

    Sigh! SPAM reported.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column.
    "If you think its expensive to hire a professional to do the job, wait until you hire an amateur."--Red Adair
    "Change is inevitable... change for the better is not."
    When you put the right degree of spin on it, the number 3|8 is also a glyph that describes the nature of a DBAs job. 😉

    Helpful Links:
    How to post code problems
    Create a Tally Function (fnTally)

  • Neha05

    Default port

    Points: 1494

    Very nice article.

Viewing 15 posts - 1 through 15 (of 25 total)

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