SQL

What it means

SQL is the language for working with data in a relational database: selecting rows by condition, joining tables, computing totals and averages, inserting and updating records.

What matters to a business is the consequence: data in a database can be pulled in any combination without anyone retyping it into a spreadsheet. That is exactly what separates record-keeping from information.

SQL describes the result, not the procedure. In a spreadsheet a calculation is dragged from cell to cell and depends on who edited the sheet and how; in SQL you state which rows and totals you want, and the database decides for itself how to get them out of millions of records. That is why the same query gives a comparable result from fresh data every month and can feed a dashboard.

How it works in practice

In accounting and warehouse systems SQL sits under every report. Where the system allows a custom query or an export over an API, tailored reporting can be built from it without changing the software.

With more data and several sources, a data warehouse and business intelligence tools go on top - and SQL stays the language in which the metrics themselves are defined.

For larger clients we run queries over the export from the accounting system to check what a standard report will not show: duplicate invoice numbers from one supplier, the same IBAN on different partners, or payments entered at the weekend. It is a cheap addition to internal control that runs all year, not only at audit time.

What to watch out for

Write access to the production database belongs to nobody who does not need it: one wrong statement can silently change every row. Analytical queries therefore run on a copy, with read-only rights.

Two queries answering the same question can give different figures: one counts revenue with credit notes, the other without, one by issue date, the other by delivery date. Until the definition is written down in one place, the meeting circles around which number is right instead of what it means.