GitHub - recursivefunk/pg-gen: Use ES6 Generators to paginate through large Postgres result sets (original) (raw)

pg-gen

Use ES6 Generators to paginate through large Postgres result sets You'll need to use Node 4.2.0 and up

Explaination

pg-gen uses a postgresql cursor for queries so that an entire result set is not returned and so memory is conserved. You utilize a generator to fetch the next n rows until there are no more results.

Why not stream the result set?

Streams are great assuming you are ready to receive said results. Otherwise, you'd have to manually manage flow using pause/resume. And this will still eventually max out the internal buffer. Generators actually pause function execution. There is no buffer, there is only ever the data you request. If streams push data, generators allow you to pull data. I advise you to use whichever is best for solving your particular problem.

Install

npm install pg-gen --save

Usage

const PgGen = require( 'pg-gen' ); const pg = require( 'pg' ); const conn = process.env.DATABASE_URL; const client = new pg.Client( conn );

// assume the necessary pg client connection stuff here ....

const pgGen = PgGen( { pg: client } ); const query = select * from table_with_so_much_data; const gen = pgGen.atMost( 10 ).lazyQuery( query );

gen().then( ( result ) => {

console.log(result); // an array with at most, 10 items

gen().then( ( result ) => {
  console.log(result); // an array with the NEXT 10 items
});

// Do this until result has a length of 0

});

API

Each generator execution will return no more than the specified number of items. It may return less upon reaching the end of the result set. A return of 0 means there is no more data.

lazyQuery([query=string], [params=object])

Tests

You'll need to have a (running) database and a table with data in order to run the tests. Create your test environment like so

$ echo "DATABASE_URL=postgres://user:$ecret@localhost/test_db" >> test/test.env
$ echo "TEST_TABLE=test_table" >> test/test.env

Then