Clean and protect¶
Finding what is wrong with a table, fixing it, and making it safe to share.
Each row is one job, with where to find it in the app, on the command line and as an MCP tool for AI agents. A dash means that surface does not offer it.
| I want to... | In the app | Command line | MCP tool |
|---|---|---|---|
| Get clean-up suggestions | Clean-up Suggestions | - | - |
| Find and drop duplicate rows | Find duplicates | --dedupe |
find_duplicates, drop_duplicates |
| Find near-duplicates | Find Near-Duplicates | - | fuzzy_duplicates |
| Fill missing values | Fill Missing Values | --impute |
fill_missing |
| Find outliers | Detect Outliers | --outliers |
detect_outliers |
| Check values against rules | Data Validation | --check |
check_rules |
| Get a data quality report | Data Quality Report | - | - |
| Spot values typed in the wrong format | Value Shapes | --shapes |
value_shapes |
| Find personal data | Detect PII | --detect-pii |
detect_pii |
| Anonymise columns | Anonymise Columns | --anonymize |
anonymize |
| Hash columns | Hash Columns | --hash-columns |
hash_columns |
| Make test data shaped like the real thing | Test Data | --test-data |
generate_test_data |