Grouping by records that are already grouped.

  • I have results from a query that have anywhere from 1-4 results. I have a subid of 1-4 which is grouped by a certain criteria. so for example...

    id subid text

    1 processing A records

    2 1000 records processing

    3 importing A records

    4 1000 records processed

    1 processing B records

    2 500 records processing

    3 importing B records

    4 1000 records processed

    Here is what my desired output is giving each group of subids 1-4 an id to be grouped together.

    id subid text

    1 1 processing A records

    1 2 1000 records processing

    1 3 importing A records

    1 4 1000 records processed

    2 1 processing B records

    2 2 500 records processing

    2 3 importing B records

    2 4 1000 records processed

  • brody.rathman (7/27/2015)


    I have results from a query that have anywhere from 1-4 results. I have a subid of 1-4 which is grouped by a certain criteria. so for example...

    id subid text

    1 processing A records

    2 1000 records processing

    3 importing A records

    4 1000 records processed

    1 processing B records

    2 500 records processing

    3 importing B records

    4 1000 records processed

    Here is what my desired output is giving each group of subids 1-4 an id to be grouped together.

    id subid text

    1 1 processing A records

    1 2 1000 records processing

    1 3 importing A records

    1 4 1000 records processed

    2 1 processing B records

    2 2 500 records processing

    2 3 importing B records

    2 4 1000 records processed

    It's very helpful to those trying to help you out if you help us first, by supplying the SQL statements necessary to create tables with the relevant structures and inserting sample data. See the first link in my signature for how to do this.

    For what you gave us, this would be:

    DECLARE @test-2 TABLE (RowID INTEGER IDENTITY PRIMARY KEY CLUSTERED, subid INTEGER, [text] VARCHAR(250));

    INSERT INTO @test-2 (subid, text)

    VALUES (1,'processing A records' ),

    (2,'1000 records processing' ),

    (3,'importing A records' ),

    (4,'1000 records processed' ),

    (1,'processing B records' ),

    (2,'500 records processing' ),

    (3,'importing B records' ),

    (4,'1000 records processed' );

    Now, to solve your issue:

    WITH cte AS

    (

    SELECT RowID, subid, [text], ROWID - subid AS Grp

    FROM @test-2

    ), cte2 AS

    (

    SELECT DISTINCT Grp

    FROM cte

    ), cte3 AS

    (

    SELECT *, RANK() OVER (ORDER BY Grp) AS ID

    FROM cte2

    )

    SELECT cte3.ID,

    subid,

    [text]

    FROM cte

    JOIN cte3 ON cte.Grp = cte3.Grp

    ORDER BY cte3.ID, subid;

    What this code does is to create "data islands", where we use the sequential number RowID, and subtract from it the group number. This will, for each group, give us a constant value (in the Grp column).

    Next, we need to get each of these values in a distinct list, and then utilize the RANK() function to assign them a sequential number. This gives us the ID value.

    Finally, we take the data islands, and join it to the CTE (common table expression) that gives us the ID value, to show all of them.

    In the above example, the results are:

    ID subid text

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

    1 1 processing A records

    1 2 1000 records processing

    1 3 importing A records

    1 4 1000 records processed

    2 1 processing B records

    2 2 500 records processing

    2 3 importing B records

    2 4 1000 records processed

    Wayne
    Microsoft Certified Master: SQL Server 2008
    Author - SQL Server T-SQL Recipes


    If you can't explain to another person how the code that you're copying from the internet works, then DON'T USE IT on a production system! After all, you will be the one supporting it!
    Links:
    For better assistance in answering your questions
    Performance Problems
    Common date/time routines
    Understanding and Using APPLY Part 1 & Part 2

Viewing 2 posts - 1 through 1 (of 1 total)

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