Checking data in the project database

Run a scan before working on a copy

A database moved from production may contain customer data, including addresses, phone numbers and identifiers. The scan helps you check whether such data is present before you give an agent access to the project. Run it with the button next to the project or with the command:

contextadmin scan --project sklep-klienta

The operation reads the database without writing any changes to it.

Review the type and location of findings

The report shows the data type, table, column and number of rows. It also shows columns not covered by the anonymization profile. Based on this, you can decide which rules need to be added.

How detection works

Detectors use checksums, plus additional conditions where the data format allows. In the measurement described, on random 11-digit strings, the checksum of PESEL (the Polish national identification number) alone produced 9.8% false positives. After adding a date validity check, the result was 0.54%. This is the result of a specific measurement, not a guarantee for any database.

An additional sample is taken in the areas where the first pass detected data. When interpreting the results, take sampling and your own application’s schema into account.

Result saved with the project

The data status in the header comes from the last scan. It is not a manually set declaration. After anonymization, run the measurement again to check the result of the changes.