Anonymize command-line reference - classify
Published 14 March 2025
Automatically discover which columns of a database contain PII (personally identifiable information).
rganonymize classify --database-engine SqlServer --connection-string "[connection string]" --classification-file classification.json
| Option | Required? | Notes |
|---|---|---|
--database-engine |  | The type of the database to classify. Must be one of:
|
--connection-string | | The connection string of the database to classify. Can also be set via a |
--classification-file | | The file path to save the classification to, including file name and file extension. |
--options-file |  | The file path to an options file. |
--scan-type | | The mode of scanning to use for classification. AI Classify finds sensitive data that rule-based classification misses, by running a local AI model to examine actual data from the database, not just column names and metadata. No data leaves your environment. See AI Classify for more details. Must be one of:
|
--connection-timeout | | The timeout (in seconds) to use when connecting to the database. Default: 90s |
--command-timeout | | The timeout (in seconds) to use when executing commands against the database. Default: 90s |
--output-all-columns | | Include all columns in the classification output file, including columns with no classification. |
--output-schema-file | | Output a schema file to aid with hand editing the classification file. The schema can be used with some IDEs to provide prompts and documentation on the expected structure when editing. |
--output | | The format to write any output. Can be one of:
|
| | The verbosity level to log at.
|
| | The location of the log file, or |
| | Show help and usage information. |
