Flat file read , conditionally extract data rows

  • This is urgent please.

    Use SSIS

    flat file sample content (use attachement, sampleFF),

    ABC 8585

    TRANID

    A4145

    CCA

    96

    FVA

    78

    ABO 8500

    IBD 8585

    FBD 8585

    TRANID

    A4146

    CCA

    90

    FVA

    71

    ABO 8110

    requirement .

    if a rowline contains "CA" or "VA"

    extract else no , unless it is tranID which ties the above two fields as follow:

    IDfieldName FieldvalueTranID(see attachment Sampleresult.jpg)

    1CCA96A4145

    2FVA78A4145

    3CCA90A4146

    4FVA71A4146

    Note: fieldvalues in the flat file are always on one row below the fieldnames and position starts at 7(for those fields that contains , CA,VA or TranID).

    for the sake of Performance , Cannot import data to database and manupilate DB side , file is too big, may be 12million rows and fields that are required are those with CA , VA and tranIDs.

  • use a Conditional Split Transformation on the Data Flow. Look for your data there.

    Gerald Britton, Pluralsight courses

  • You aren't going to be able to use a standard SSIS Flat File Source for this, you are going to have to do scripting to read the file row-by-row. Somthing like this psuedocode:

    Open File

    Read First Row

    IF TRANID Then read the next row and store value in variable

    Read next row if contains "CA" or "VA" store value in variable

    Read next row and store in value in variable

    Write stored values to Output

  • OK, I think I understand what you want. But what is your question?

    --Edit: Jack's on to something. Use an asynchronous Script Component for this.

    The absence of evidence is not evidence of absence.
    Martin Rees

    You can lead a horse to water, but a pencil must be lead.
    Stan Laurel

  • Jack , Would you direct me to a resource I can refer to or yourself may write a bit of code sample to that I can expand?

  • Quick suggestion as this isn't as complex as it seems at first, import the file's content into a staging table, one line per row and group/parse it from there.

    😎

    USE tempdb;

    GO

    SET NOCOUNT ON;

    IF OBJECT_ID('dbo.TBL_LINE_IMPORT') IS NOT NULL DROP TABLE dbo.TBL_LINE_IMPORT;

    CREATE TABLE dbo.TBL_LINE_IMPORT

    (

    LIMP_ID INT IDENTITY(1,1) NOT NULL CONSTRAINT PK_DBO_TBL_LINE_IMPORT_LIMP_ID PRIMARY KEY CLUSTERED

    ,LIMP_LINE VARCHAR(250) NOT NULL

    );

    /* The file content imported line by line */

    INSERT INTO dbo.TBL_LINE_IMPORT (LIMP_LINE)

    VALUES

    ('ABC 8585')

    ,('TRANID ')

    ,(' A4145')

    ,('CCA ')

    ,(' 96')

    ,('FVA ')

    ,(' 78')

    ,('ABO 8500')

    ,('ABD 8585')

    ,('TRANID ')

    ,(' A4146')

    ,('CCA ')

    ,(' 90')

    ,('FVA ')

    ,(' 71')

    ,('ABO 8110');

    ;WITH BASE_GROUP AS

    (

    SELECT

    LI.LIMP_ID

    FROM dbo.TBL_LINE_IMPORT LI

    WHERE LTRIM(RTRIM(LI.LIMP_LINE)) = 'TRANID'

    UNION ALL

    SELECT MAX(LI.LIMP_ID) + 1 AS LIMP_ID

    FROM dbo.TBL_LINE_IMPORT LI

    )

    ,NUMBERED_GROUP AS

    (

    SELECT

    ROW_NUMBER() OVER

    (

    PARTITION BY (SELECT NULL)

    ORDER BY (SELECT NULL)

    ) AS BG_RID

    ,BG.LIMP_ID

    FROM BASE_GROUP BG

    )

    ,PARSE_GROUP AS

    (

    SELECT

    NG.BG_RID

    ,NG.LIMP_ID AS START_ID

    ,LG.LIMP_ID - 1 AS END_ID

    FROM NUMBERED_GROUP NG

    INNER JOIN NUMBERED_GROUP LG

    ON NG.BG_RID = LG.BG_RID - 1

    )

    ,COMBINED_SET AS

    (

    SELECT

    LI.LIMP_ID

    ,PG.BG_RID

    ,LTRIM(RTRIM(LI.LIMP_LINE)) AS LI_LINE

    ,LTRIM(RTRIM(L2.LIMP_LINE)) AS L2_LINE

    FROM dbo.TBL_LINE_IMPORT LI

    LEFT OUTER JOIN dbo.TBL_LINE_IMPORT L2

    ON LI.LIMP_ID = (L2.LIMP_ID - 1)

    CROSS APPLY PARSE_GROUP PG

    WHERE LI.LIMP_ID BETWEEN PG.START_ID AND PG.END_ID

    )

    ,FINAL_SET AS

    (

    SELECT

    CASE

    WHEN CS.LI_LINE = 'CCA' THEN CS.LI_LINE

    WHEN CS.LI_LINE = 'FVA' THEN CS.LI_LINE

    END AS fieldName

    ,CASE

    WHEN CS.LI_LINE = 'CCA' THEN CS.L2_LINE

    WHEN CS.LI_LINE = 'FVA' THEN CS.L2_LINE

    END AS Fieldvalue

    ,MAX(CASE WHEN CS.LI_LINE = 'TRANID' THEN CS.L2_LINE END) OVER (PARTITION BY CS.BG_RID) AS TranID

    FROM COMBINED_SET CS

    )

    SELECT

    FS.fieldName

    ,FS.Fieldvalue

    ,FS.TranID

    FROM FINAL_SET FS

    WHERE FS.fieldName IS NOT NULL;

    Results

    fieldName Fieldvalue TranID

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

    CCA 96 A4145

    FVA 78 A4145

    CCA 90 A4146

    FVA 71 A4146

Viewing 6 posts - 1 through 5 (of 5 total)

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