Maintaining an InfoSource (flat file) 
Purpose
You can load data from flat files (CSV or ASCII files) into BW.
You can load the following data types:
Transaction Data
Master Data
-
Hierarchies
Prerequisites
Please note the following regarding Excel CSV files:
Excel files use delimiters to separate fields. In the European version, the delimiter ";" is used. You can also use other delimiters. You must specify the delimiter used in the Scheduler.
Fields that are not filled in an Excel CSV file are filled with a blank character in the case of field characters.
If the delimiters have been used inconsistently in an Excel CSV file, the "wrong" delimiter is read as a character, and both fields are brought together and possibly cut short. The following fields are then no longer in the correct order.
Please note the following regarding Excel CSV files, such as ASCII files:
- If your file contains headers that you do not want to load, enter how many headers should be ignored during loading on the tabstrip External Data Parameters in the Scheduler. This gives you the option of keeping the column headers in your file.
- Depending on the conversion exit, either enter or do not enter a zero at the beginning.
- For entering the date, use as a rule the format ‘YYYYMMDD’ without internal delimiters. Depending on the conversion exit, you can also use other formats.
- If you upload data using IDocs, make note of the 1.000 byte limit per data record length. This limit does not apply for uploading data with the PSA.
Notes on Uploading
When you upload external data, you have the option of loading the data from any workstation into BW. From a performance point of view, however, you should store the data onto an application server and load it from there into BW. In this way, you can load the data in the background.
- If you want to upload a large amount of transaction data from a flat file using the IDoc loading method, and if you have the option of specifying the file type of the flat file, you should create the flat file as an ASCII file. From a performance point of view, uploading the data from an ASCII file is the most cost-effective method. Generating an ASCII file might mean a larger workload.