Updating rows based on identical column values

  • Hi,

    I have a staging table that contains data obtained from a xml file import via SSIS.

    In instances where 'Unit' values are NULL, I need to convert these into correct values. These are based on the A_Codes. Records that contain NULL values (in Unit column) occur are where there has been a duplication of previous A_Codes (i.e. where Line_Id > 1) so there will always be a comparison to match on. Hope that makes sense!

    Therefore, in the examples below, lines (PK_Id) 10, 11 & 12 need the 'Unit' values be updated.

    PK_Id is the actual file 'Identity Specification' (incremented by 1)

    rows where Line_Id = 1 will ALWAYS have a 'Unit' value (not NULL)

    Example:

    Table A (after xml import)

    AB_Id, A_Code, Unit, Line_Id, PK_Id

    FC002, 0041, %, 1, 1

    FC002, 0043, %, 1, 2

    FC002, 0044, %, 1, 3

    FC002, 0045, %, 1, 4

    FC002, 0046, %, 1, 5

    FC002, 0047, %, 1, 6

    FC002, 0049, mg/kg, 1, 7

    FC002, 0052, mg/kg, 1, 8

    FC002, 0055, mg/kg, 1, 9

    FC002, 0044, NULL, 2, 10

    FC002, 0052, NULL, 2, 11

    FC002, 0052, NULL, 3, 12

    Table_A (after conversions)

    AB_Id, A_Code, Unit, Line_Id, PK_Id

    FC002, 0041, %, 1, 1

    FC002, 0043, %, 1, 2

    FC002, 0044, %, 1, 3

    FC002, 0045, %, 1, 4

    FC002, 0046, %, 1, 5

    FC002, 0047, %, 1, 6

    FC002, 0049, mg/kg, 1, 7

    FC002, 0052, mg/kg, 1, 8

    FC002, 0055, mg/kg, 1, 9

    FC002, 0044, %, 2, 10

    FC002, 0052, mg/kg, 2, 11

    FC002, 0052, mg/kg, 3, 12

    Notes:

    In the above example the 10th record (Line_Id = 10) has had the 'Unit' value changed from NULL to % (matched on row 3).

    In the above example the 11th record (Line_Id = 11) has had the 'Unit' value changed from NULL to mg/kg (matched on row 8).

    In the above example the 12th record (Line_Id = 12) has had the 'Unit' value changed from NULL to mg/kg (matched on row 8).

    Any ideas please?

    Thanks in advance,

  • Try the following code,

    DECLARE @Table_A TABLE (AB_Id nvarchar(5), A_Code nvarchar(5), Unit nvarchar(5), Line_Id int, PK_Id int)

    INSERT @Table_A

    SELECT 'FC002', '0041', '%', 1, 1

    UNION SELECT 'FC002', '0043', '%', 1, 2

    UNION SELECT 'FC002', '0044', '%', 1, 3

    UNION SELECT 'FC002', '0045', '%', 1, 4

    UNION SELECT 'FC002', '0046', '%', 1, 5

    UNION SELECT 'FC002', '0047', '%', 1, 6

    UNION SELECT 'FC002', '0049', 'mg/kg', 1, 7

    UNION SELECT 'FC002', '0052', 'mg/kg', 1, 8

    UNION SELECT 'FC002', '0055', 'mg/kg', 1, 9

    UNION SELECT 'FC002', '0044', NULL, 2, 10

    UNION SELECT 'FC002', '0052', NULL, 2, 11

    UNION SELECT 'FC002', '0052', NULL, 3, 12;

    WITH Conversion (Unit, A_Code) AS (

    SELECTDISTINCT

    Unit, A_Code

    FROM@Table_A

    WHERELine_Id = 1

    )

    SELECTA.AB_Id,

    A.A_Code,

    COALESCE(A.Unit, C.Unit) [Unit],

    Line_Id,

    PK_Id

    FROM@Table_A A LEFT OUTER JOIN

    Conversion C ON A.A_Code = C.A_Code

  • or:

    -- *** Test Data ***

    DECLARE @a TABLE

    (

    AB_Id char(5) NOT NULL

    ,A_Code char(4) NOT NULL

    ,Unit varchar(10) NULL

    ,Line_Id int NOT NULL

    ,PK_Id int NOT NULL

    )

    INSERT INTO @a

    SELECT 'FC002', '0041', '%', 1, 1

    UNION ALL SELECT 'FC002', '0043', '%', 1, 2

    UNION ALL SELECT 'FC002', '0044', '%', 1, 3

    UNION ALL SELECT 'FC002', '0045', '%', 1, 4

    UNION ALL SELECT 'FC002', '0046', '%', 1, 5

    UNION ALL SELECT 'FC002', '0047', '%', 1, 6

    UNION ALL SELECT 'FC002', '0049', 'mg/kg', 1, 7

    UNION ALL SELECT 'FC002', '0052', 'mg/kg', 1, 8

    UNION ALL SELECT 'FC002', '0055', 'mg/kg', 1, 9

    UNION ALL SELECT 'FC002', '0044', NULL, 2, 10

    UNION ALL SELECT 'FC002', '0052', NULL, 2, 11

    UNION ALL SELECT 'FC002', '0052', NULL, 3, 12

    -- *** End Test Data ***

    UPDATE A1

    SET Unit = A2.Unit

    FROM @a A1

    JOIN @a A2

    ON A1.A_Code = A2.A_Code

    AND A2.Line_Id = 1

    WHERE A1.Unit IS NULL

    SELECT *

    FROM @a

  • This was removed by the editor as SPAM

  • Excellent,

    Thanks to all!

Viewing 5 posts - 1 through 4 (of 4 total)

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