17
votes

Does any one have a tutorial on how to import a fixed width flat file into a database using an SSIS package?

I have a flat file containing columns with varying lengths.

Column name    Width
-----------    -----
First name        25
Last name         25
Id                 9
Date               8

How do I convert a flat file into columns?

3

3 Answers

55
votes

Here is a sample package created using SSIS 2008 R2 that explains how to import a flat file into a database table.

  • Create a fixed-width flat file named Fixed_Width_File.txt with data as shown in the screenshot. The screenshot uses Notepad++ to display the file contents. It has the capability to show the special characters like carriage return and line feed. CR LF denotes the row delimiters Carriage return and Line feed.

Flat file data

  • In the SQL server database, create a table named dbo.FlatFile using the create script provided under SQL Scripts section.

  • Create a new SSIS package and add a new OLE DB Connection manager that would connect to the SQL Server database. Let's assume that the OLE DB Connection manager is named as SQLServer.

Connection manager

  • On the package's control flow tab, place a Data Flow Task.

Data flow task

  • Double-click on the data flow task and you will be taken to the data flow tab. On the data flow tab, place a Flat File Source. Double-click on the flat file source and the Flat File Source Editor will appear. Click the New button to open the Flat File Connection Manager Editor.

  • On the General section of the Flat File Source Editor, enter a value in Connection manager name (say Source) and browse to the flat file location and select the file. This example uses the sample file in the path C:\temp\Fixed_Width_File.txt If you have header rows in your file, you can enter a value 1 in the Header rows to skip textbox to skip the header row.

Flat file connection manager editor General

  • Click on the Columns section. Change the font according to your choice I chose Courier New so I could see more data with less scrolling. Enter the value 69 in the Row width text box. This value is the sum of width of all your columns + 2 for the row delimiter. Once you have set the correct row width, you should see the fixed width file data correctly on the Source data columns section. Now, you have to click at the appropriate locations to determine the column limits. Note the sections 4, 5, 6 and in the below screenshot.

Flat file connection manager editor Columns

  • Click on the Advanced section. You will notice 5 columns created for you automatically based on the column limits that we set on the Columns section in the previous step. The fifth column is for row delimiter.

Flat file connection manager editor Advanced

  • Rename the column names as FirstName, LastName, Id, Date and RowDelimiter

Flat file connection manager editor Advanced Renamed

  • By default, the columns will be set with DataType string [DT_STR]. If we are fairly certain, that a certain column will be of different data type, we can configure it in the Advanced section. We will change Id column to be of data type four-byte signed integer [DT_I4] and Date column to be of data type date [DT_DATE]

Flat file connection manager editor Advanced Id column

Flat file connection manager editor Advanced Date column

  • Click on the Preview section. The data will be shown as per the column configuration.

Flat file connection manager editor Preview

  • Click OK on the Flat file connection manager editor and the flat file connection will be assigned to the Flat File Source in the data flow task.

Flat file editor connection

  • On the Flat File Source Editor, click on the Columns section. You will notice the columns that were configured in the flat file connection manager. Uncheck the RowDelimiter because we won't need that.

Flat file editor columns

  • On the data flow task, place an OLE DB Destination. Connect the output from the Flat file source to the OLE DB Destination.

Data flow task

  • On the OLE DB Destination Editor, select the OLE DB Connection manager named SQLServer and set the Name of the table or the view drop down to [dbo].[FlatFile]

OLE DB Destination connection

  • On the OLE DB Destination Editor, click on the Mappings section. Since the column names in the flat file connection manager are same as the columns in the database, the mapping will take place automatically. If the names are different, you have to manually map the columns. Click OK.

OLE DB Destination columns

  • Now the package is ready. Execute the package to load the fixed-width flat file data into the database.

Package execution

  • If you query the table dbo.FlatFile in the database, you will notice the flat file data imported into the database.

Data imported into table

This sample should give you an idea about how to import fixed-width flat file into database. It doesn't explain how to handle error logging but this should get you started and help you discover other SSIS related features when you play with packages.

Hope that helps.

SQL Scripts:

CREATE TABLE [dbo].[FlatFile](
    [Id] [int] NOT NULL,
    [FirstName] [varchar](25) NOT NULL,
    [LastName] [varchar](25) NOT NULL,
    [Date] [datetime] NOT NULL
)
1
votes

In the derived column transformation you can use SUBSTRING() function for each of the column. Example:

Columns DerivedColumn

FirstName SUBSTRING(Data, startFrom, Length);

Here the FirstName has width 25 so if we consider that from the 0th position then in the derived column you should specify it by giving SUBSTRING(Data, 0, 25);

Similarly for other columns.

1
votes

Very well explained, Siva! Your tutorial and excellent illustrations point out what Microsoft should have made clear

  1. that the width for a fixed length row has to include the Carriage Return and Line Feed (CR & LF) characters (which I figured out because the preview showed the rows were not lining up correctly)
  2. the all important step of defining an extra column to contain those CR & LF characters, even though they won't be imported. I figured this out, too. I would have benefited by finding your answer before I began.

Without those two things, an attempt to run the import will give this error message: The data conversion for column "Column x" returned status value 4 and status text "Text was truncated or one or more characters had no match in the target code page.".

I have added in this error text in hopes someone will find this page while searching for the cause of their error. Your turorial is worth finding, even if after the fact!