README

pg-anonymizer

Export your PostgreSQL database anonymized. Replace all sensitive data thanks to faker. Output to a file that you can easily import with psql.

Usage

Run this command by giving a connexion string and an output file name (no need to install first thanks to npx):

npx pg-anonymizer postgres://user:secret@localhost:1234/mydb -o dump.sql

☝️ This command requires pg_dump. It may already be installed as soon as PostgreSQL is installed.

Specify list of columns to anonymize

Use --list option with a comma separated list of column name:

npx pg-anonymizer postgres://localhost/mydb \
  --list=email,firstName,lastName,phone

Specifying another list via --list replace the default automatically anonymized values:

email,name,description,address,city,country,phone,comment,birthdate

You can also specify the table for a column using the dot notation:

public.user.email,public.product.description,email,name

Customize replacements

You can also choose which faker function you want to use to replace data (default is faker.random.word):

npx pg-anonymizer postgres://localhost/mydb \
  --list=firstName:faker.name.firstName,lastName:faker.name.lastName

:point_right: You don't need to specify faker function since the command will try to find correct function via column name.

You can use plain text too for static replacements:

npx pg-anonymizer postgres://localhost/mydb \
  --list=textcol:hello,jsoncol:{},intcol:12

You can even use your custom replacements function from your own javascript module. Here is a simple example to mask all the email.

npx pg-anonymizer postgres://localhost/mydb \
  --extension ./myExtension.js \
  --list=email:extension.maskEmail

// myExtension.js
module.exports = {
  maskEmail: (email) => {
   const [name, domain] = email.split('@');
   const { length: len } = name;
   const maskedName = name[0] + '...' + name[len - 1];
   const maskedEmail = maskedName + '@' + domain;
   return maskedEmail;
  }
};

Locale (i18n)

Use -l to change the locale used by faker (default: en)

Import the anonymized file

The anonymized output file is plain SQL text, you can import it with psql.

psql -d mylocaldb < output.sql

Why

There are a bunch of competitors, still I failed to use them:

postgresql_anonymizer may be hard to setup and may be cumbersome for simple usage. Still, I guess it's the best solution.
pganonymize fails when it does not use public schema or columns have uppercase characters
pganonymizer also fails with simple cases. Errors are not explicit and silent.

Usage no npm install needed!