pg-anonymizer

0.8.1 verified Sat Apr 25 auth: no javascript

pg-anonymizer is a CLI tool (built with oclif) that exports a PostgreSQL database and anonymizes sensitive columns using faker.js. It wraps pg_dump and replaces data on the fly. Current stable version 0.8.1, released Feb 2024, with ~5 releases per year. Key differentiators: no database installation, works as a pipe, supports custom faker functions, extension files, and config files. Alternative to pg_dump with custom scripts or postgresql_anonymizer, but simpler and portable.

Common errors

error pg_dump: command not found ↓

cause pg_dump is not installed on the system.

fix

Install PostgreSQL client tools (e.g., on Ubuntu: sudo apt-get install postgresql-client).

error Error: Cannot find module 'pg-anonymizer' ↓

cause When using require in an older CJS project without transpilation, but package is ESM-only in some versions.

fix

Use dynamic import() or switch to ESM: 'import pgAnonymizer from "pg-anonymizer"'.

error Error: Column 'nonexistent' does not exist in any table ↓

cause Specified a column name that does not exist in any database table.

fix

Check the column name spelling and ensure it exists in the database.

Warnings

gotcha pg_dump must be installed separately; it is not bundled with npm package. ↓

fix Install PostgreSQL client tools (e.g., apt-get install postgresql-client) or ensure pg_dump is in PATH.

breaking v0.6.0 changed default faker from faker.random.word to auto-detection based on column name; custom --columns without faker function now auto-detects. ↓

fix If you relied on faker.random.word default, specify columns explicitly with ':faker.random.word'.

deprecated The --columns option with dot-notation for tables (e.g., public.user.email) may be deprecated in future; prefer using --skip for tables. ↓

fix Use --skip to exclude tables from anonymization instead of prefixing column names.

gotcha Unexpected anonymization of non-sensitive columns: if no --columns specified, a default list of columns (email, name, etc.) is anonymized, potentially including non-sensitive data. ↓

fix Always specify --columns to avoid accidentally anonymizing unintended columns.

gotcha Output file path is relative to current working directory, not the script location. ↓

fix Use absolute paths for --output to avoid confusion.

deprecated Node.js >=12 required; Node 10 support dropped in v0.6.0. ↓

fix Upgrade Node.js to >=12.

Install

npm install pg-anonymizer

yarn add pg-anonymizer

pnpm add pg-anonymizer

Imports

default
wrong
```
const pgAnonymizer = require('pg-anonymizer')
```
correct
```
import pgAnonymizer from 'pg-anonymizer'
```
ESM and CJS both supported (dual package). The default export is the CLI runner.

anonymize

wrong

const { anonymize } = require('pg-anonymizer')

correct

import { anonymize } from 'pg-anonymizer'

Named export for programmatic usage; available since v0.6.

types
```
import type { Config } from 'pg-anonymizer'
```
TypeScript types are bundled; use import type for type-only imports.

Quickstart

Anonymize a PostgreSQL database export by specifying columns to anonymize, using npx without installation.

npx pg-anonymizer postgres://user:pass@localhost:5432/mydb --columns=email,firstName,lastName --output dump.sql