Integrating with SAP and systems that don't want to talk

If someone in your company opens SAP every morning, exports a report like VA05, saves an Excel file and then re-types it into another system by hand — you are not alone. SAP and similar large systems rarely talk willingly to their surroundings. But that does not mean you are doomed to manual re-typing. It only means the integration should be built around what SAP actually gives you: periodic exports.

The export as an input, not an obstacle

You do not always need a direct connection to SAP to have a reliable data flow. Very often it is enough that the export you already make anyway — an Excel or text file — is automatically picked up, processed and loaded into the system you need. A person presses the export button once, or the export is scheduled, and everything else runs on its own.

The key is to treat the file not as a final result, but as an input to a process that checks it, cleans it and matches it against existing data.

Millions of rows are not a problem if processed properly

A SAP export can be huge — over a million rows is not unusual. Excel can barely open that, and a person will never review it honestly. That is why such files are not processed by hand, but in controlled steps:

• matching — every row is linked to the right customer, order or item by a key;
• deduplication — the same record appearing in several exports enters only once;
• validation — rows with missing or nonsensical data are set aside instead of quietly corrupting the database;
• update, not duplicate — a newer export refreshes the existing record rather than creating a new one.

When this is done right, a million rows are processed in a few minutes and unattended, instead of someone spending half a day in Excel.

Who wins when exports differ

With periodic exports there is always overlap — the same sales document appears in several files, in different versions. There must be a clear rule about which version is authoritative. Usually it is the newest file that contains that document, but the rule has to be written down and consistent, otherwise you get data that changes with no logic at all.

Data you do not trust is worse than data you do not have.

From manual work to a flow you trust

The goal is not only to eliminate re-typing. The goal is for the data flow to become so reliable that no one thinks about it anymore — you know when the last export was processed, how many rows entered, how many were rejected and why. Only then does data from SAP become something you can rely on for planning and reporting.

If your company also runs on SAP exports that someone re-types by hand, that can most likely be turned into an automatic, verifiable flow. We will gladly look at your files and suggest how.