SSIS - Write table to multiple files

Question

Post reply

SSIS - Write table to multiple files

darkangelBDF

Old Hand

Points: 349

September 1, 2021 at 12:16 pm

#3925803

Hi there, everyone.

I have an SSIS package where I have to write a large table to smaller csv files. I have created a stored procedure as below that is called by an OleDb Source inside a workflow task. This then needs to write each result to a file. I'm very confused as to how to do this. Below is the stored procedure. Can anyone steer me in the right direction as to how to approach this, please?

SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO

CREATE PROCEDURE [dbo].[GEO_FileExtract]

AS

BEGIN TRY

DECLARE @TempTable TABLE (
[IDNO] [varchar](16) NULL,
[GEO001] [varchar](15) NULL,
[GEO002] [varchar](5) NULL,
[GEO003] [varchar](5) NULL,
[GEO004] [varchar](5) NULL,
[GEO005] [varchar](5) NULL)


WHILE EXISTS (
SELECT TOP 1 *
FROM dbo.GEO AS INXG
WHERE IsProcessed = 0)

BEGIN

INSERT INTO @TempTable (
IDNO,
GEO001,
GEO002,
GEO003,
GEO004,
GEO005)
SELECT TOP 50000
INXG.IDNO,
INXG.GEO001,
INXG.GEO002,
INXG.GEO003,
INXG.GEO004,
INXG.GEO005
FROM dbo.GEO AS INXG
ORDER BY INXG.IDNO

UPDATE INXG
SET IsProcessed = 1
FROM dbo.GEO AS INXG
JOIN @TempTable AS TT
ON INXG.IDNO = TT.IDNO
WHERE INXG.IsProcessed = 0

SELECT
IDNO AS IDENTIFIERNUMBER,
TT.GEO001 AS EACODEOFCONSUMER,
TT.GEO002 AS PROVINCEOFCONSUMER,
TT.GEO003 AS DISTRICTOFCONSUMER,
TT.GEO004 AS METROPOLITANOFCONSUMER,
TT.GEO005 AS MAINPLACEOFCONSUMER
FROM @TempTable AS TT

END

END TRY

BEGIN CATCH

SELECT
    ERROR_NUMBER() AS ErrorNumber,
    ERROR_STATE() AS ErrorState,
    ERROR_SEVERITY() AS ErrorSeverity,
    ERROR_PROCEDURE() AS ErrorProcedure,
    ERROR_LINE() AS ErrorLine,
    ERROR_MESSAGE() AS ErrorMessage;

END CATCH

September 6, 2021 at 11:52 am

This was removed by the editor as SPAM
September 21, 2021 at 10:28 am

This was removed by the editor as SPAM

Viewing 12 posts - 1 through 11 (of 11 total)

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

homebrew01 SSC Guru Points: 55454 More actions · Answer 1

homebrew01

SSC Guru

Points: 55454

September 1, 2021 at 12:58 pm

#3925829

BCP aka Bulk Import/Export is 1 option

Phil Parkin SSC Guru Points: 245813 More actions · Answer 2

I have a few quick comments.

Using a table variable for such large volumes of data is usually not a great idea. A temp table will almost certainly perform better.
Shouldn't your initial INSERT include "WHERE IsProcessed = 0" to avoid reselecting rows already selected?
By doing a SELECT in your CATCH block, you are telling the proc not to fail on error, but merely to return a resultset. Is that what you want? If you want it to properly fail on error, put a THROW or RAISERROR in there.
Use an OLEDB data source to execute the proc and write to a file.
Once the file is generated, use a script task or file system tasks to copy it to your various desired names/locations.

If you haven't even tried to resolve your issue, please don't expect the hard-working volunteers here to waste their time providing links to answers which you could easily have found yourself.

Brahmanand Shukla SSC Eights! Points: 968 More actions · Answer 3

Do you know the number of files (smaller files) to be created out of the table? Or it's dynamic?

SQL Server Carpenter - Consultant & Trainer
Author of Learn T-SQL from Scratch

darkangelBDF Old Hand Points: 349 More actions · Answer 4

Hi there. Yes, it does include a where IsProcessed =! Or should. Seems I forgot to put it back. I’ll use a temp table instead. Thanks for the advice.

darkangelBDF Old Hand Points: 349 More actions · Answer 5

Hi there, Brahmanand. The row count is dynamic as it changes. The max size per file should be around 100MB. This will be sent to Azure BLOB and copied to Snowflake.

Brahmanand Shukla SSC Eights! Points: 968 More actions · Answer 6

I recommend the following steps:

Identify the total row count and keep it in a variable [A]
Identify the average row size (in bytes), and derive the average number of rows that will fit in the 100 MB, and keep it in a variable
Divide A with B (A/B), apply the ceiling function to get the nearest integer [C]. This will become the number of groups.
Use NTILE() function on the original table and supply the variable (C) to the NTILE function. Overall rows will be divided into groups.
Each individual group is a 100 MB equivalent data/rows.
In SSIS you can use a loop container and iterate from 1 to the number of groups (C). Store the iteration number/group number in an SSIS variable.
Add a Data Flow task in the loop container
1. The Source shall be OLEDB/ADO NET source. Write the query in the source and filter the data by Group Number.
2. The Destination shall be a Flat-File destination.

You can refer to the following articles for dividing a dataset into equal groups.

Divide the rows in equals batches

Divide the rows of a column in equals batches in grid format

SQL Server Carpenter - Consultant & Trainer
Author of Learn T-SQL from Scratch

Phil Parkin SSC Guru Points: 245813 More actions · Answer 7

I misunderstood the requirement – you want to split the rows contained in the table across a number of files. My comment (5) therefore makes no sense whatsoever!

The method described above looks like it will work. Another possibility is that you output to a single large file and then run a process afterwards which splits that file into multiple files.

If you haven't even tried to resolve your issue, please don't expect the hard-working volunteers here to waste their time providing links to answers which you could easily have found yourself.

darkangelBDF Old Hand Points: 349 More actions · Answer 8

darkangelBDF

Old Hand

Points: 349

September 4, 2021 at 6:40 am

#3926799

Wow, awesome. Thank you!

Jeff Moden SSC Guru Points: 1000539 More actions · Answer 9

matilda wrote:

Thanks for this case. Very interesting solution!

Sounds like a typical generic post as a precursor to SPAM. "Santa is watching"!

--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.
"Change is inevitable... change for the better is not".

Helpful Links:
How to post code problems
How to Post Performance Problems
Create a Tally Function (fnTally)
Intro to Tally Tables and Functions