pgmock2

An NPM module for mocking a connection to a PostgreSQL database.

The module mocks a pg module connection to a PostgreSQL database. Both the pg.Client and pg.Pool classes have a query method, therefore the mock connection can be used to simulate an instance of either class.

Documentation

Live documentation: here

To (re)generate documentation: npm run docs

Installation

Installation via npm.

npm i --dev-save pgmock2

Usage

The idea is to simulate a connection to a database. To enable that simulation, we need to first add data.

Adding Queries and their Responses

// Simple type checking validation.
const PgMock2 = require('./pgmock2').default;
const client = new PgMock2();

client.add('SELECT * FROM employees where id=$1', ['number'], {
    rowCount: 1,
    rows: [
        { id: 1, name: 'John Smith', position: 'application developer' }
    ]
});

Using TypeScript

import PgMock2 from 'pgmock2';
const pg = new PgMock2();

pg.add('SELECT * FROM employees where id=$1', ['number'], {
    rowCount: 1,
    rows: [
        { id: 1, name: 'John Smith', position: 'application developer' }
    ]
});

(async function() {
    const clinet = await pg.connect();
})();

Parameters of the add Method

Query

The first parameter of the add method is the query we add to the mock DB.

Later, we can use a mock connection to retrieve a response to this query. Internally, the query is normalized (disregards whitespace and is made case insensitive).

Values Validation

The second parameter is an array used to validate any values passed with the query. In the example above, the $1 requires a value. In the validation array, we pass the string number.

Since the validation criterion is a string, the only valid values that can be used at query time must be typeof "number". Functions can also be used to validate values (described later).

If a query does not require values, simply pass an empty array.

// Quering without passing values.
client.add('SELECT * FROM employees', [], {
    rowCount: 10,
    rows: [
        { id: 1, name: 'John Smith', position: 'application developer' },
        { id: 2, name: 'Jane Doe', position: 'test engineer' }
        // ... more employees omitted ...
    ]
});

Query Response

The thrid parameter is the response returned if the values supplied to to the query method were determined to be valid.

The response MUST have the same interface as a pg.QueryResponse.

Quering the Mock DB

Now we can create a mock connection and query for data.

// Get a mock db connection.
client.connect();

client.query('select * from employees where id=$1;', [1])
    .then((data) => console.log(data))
    .catch((err) => console.log(err.message));

Since the query is valid and the values passed are correct in number and type, we should see the expected output.

{ rowCount: 1,
  rows:
   [ { id: 1, name: 'John Smith', position: 'application developer' } ] }

Using Functions for Validation

For more advanced query value validation (beyond just simple type validation) we can use functions.

Let's say that our employee IDs must be whole numbers greater than 0. We can use a validation function like this:

// Advanced validation with functions.
const pgmock = require('pgmock2');
const pg = new pgmock();

const validateId = (id) => {
    return (
        typeof(id) === 'number'
        && isFinite(id)
        && id > 0
        && id === Number(id.toFixed(0)
    );
}

pg.add('SELECT * FROM employees WHERE id = $1', [validateId], {
    rowCount: 1,
    rows: [
        { id: 1, name: 'John Smith', position: 'application developer' }
    ]
});

Mocking Pool/PoolClient

To mock a pg Pool/PoolClient workflow.

import pgmock, { getPool } from 'pgmock2';
...
// Using pg instance from above examples.
const pool = getPool(pg);

(async function() {
    const client = await pool.connect();
    const res =  await client.query('select * from employees');
    
    console.log(res.rows);
    client.release();
})();

Tests

Tests are found in the test directory. To execute them, run:

npm run test

Docker Test

To run the tests in a docker environment:

npm run test:docker