The problem
Tools that work with SAP need knowledge of how the system is structured: which tables exist, which fields, and how they relate. That knowledge lives in the SAP Data Dictionary. But an ERP system has tens of thousands of tables. Requesting everything every time makes the tool slow and puts load on SAP; copying everything in one go is heavy and soon goes out of date.
The approach: fetch when needed
- On request. Information about a table is only fetched when someone works with it, and is then stored locally.
- With a shelf life. Stored knowledge expires after a while and is then refreshed, so it stays current.
- With priority. Frequently used master data first, archive tables last.
- Never everything at once. No heavy, complete copy of the system.
Knowledge that builds up
The best effect: people and AI can leave notes on every table. "This custom table holds the price agreements per customer." That way the organisation builds up knowledge of its own custom development step by step, knowledge that would otherwise only exist in people's heads.
What it gives you
- Fast tools, even on large systems.
- Little load on the SAP system.
- Growing, shared knowledge about your own data.
What to watch out for
- Store only structural information, not business data, unless that is a deliberate choice.
- Let outdated knowledge expire automatically.
- Give notes a place where everyone can find them, people and AI alike.
Would you like advice on taking load off your SAP system? Request a free advice conversation.
Updated on