Table of Contents
PostGraphile is packaged as a Node.js module, you can install it with
yarn (depending on your preference) - users tend to have fewer issues with
yarn, so we recommend it.
We recommend using the latest LTS version of Node.js and PostgreSQL, but we have limited support for older versions. Using newer released versions should work fine (we don't recommend using with alpha/beta versions though).
These aren't exactly "requirements", but they will impact your PostGraphile experience.
- Use primary keys: if you don't have primary keys on your tables then they
won't get the
nodeIdglobally unique identifier interface. Further if you don't have unique constraints then you won't be able to use the automatic update/delete mutations.
- Use foreign keys: we infer relations between tables using foreign key constraints; if you don't use these constraints then we won't know there's a relationship between the tables. There are plugins to get around this (using smart comments) but it's highly recommended that you use PostgreSQL's built in relations support.
- Don't use column-based SELECT grants: column-based grants work well for
UPDATE(especially when combined with
--no-ignore-rbac!), but they don't make sense for
DELETEand they cause issues when used with
SELECT. Quite a few things in PostGraphile depend on full-table
SELECTgrants; if you don't want to use full-table grants then you will need to disable the default mutations and use custom mutations instead (because we use
RETURNING *on the mutations), you may also have to miss out on computed columns (because we pass the entire row object to the function, though these typically still work if you're using
LANGUAGE sqlrather than
LANGUAGE plpgsql/etc). It's recommended that you instead split your tables on permission boundaries and use one-to-one relations to join them together again - this also makes writing your RBAC/RLS policies simpler. If you want to omit a column entirely then you can use the
- Function restrictions: we have pretty good support for PostgreSQL functions, but there's some common function restrictions you should check out.
- Use unique constraints rather than unique indexes when appropriate: we
use unique constraints to create mutations such as
updateUserByUsername; note that "PostgreSQL automatically creates a unique index when a unique constraint or primary key is defined for a table." -- PG docs
On top of this standard PostgreSQL best practices apply: use indexes carefully for performance, use constraints to ensure your data is valid and consistent, use triggers to take an action when something happens, etc.
From PostGraphile v4 onwards, PostGraphile requires Node.js version 8.6+ which provides
native support for
await and supports many of the ES2017 and ES2018
For best results we recommend you use the latest stable release of PostgreSQL that we officially support (v11 at time of writing), however it should run well on 9.6 or higher and anything that breaks v9.6 support will be deemed a breaking change. The absolute earliest version it will run well against is v9.4, however we do not support this version officially - we strongly recommend you upgrade.
Basic operation including introspection.
Introduces Row-Level Security - important for securing your schema.
missing_ok parameter to the
function - without this you'll need to ensure all
reference settings that always exist (e.g. you may need to set them on the
--watch is also only officially supported on 9.6+ (although it might work on
PostgreSQL 10 solves a number of performance issues - the most interesting of which for us is a significant performance boost to Row Level Security policies!
Introduces "identity columns" which we have preliminary support for.
Seems to work.
PostGraphile is developed on Mac OS X and tested on GNU/Linux. It works on Windows and we would like to officially support Windows but no-one in the core team uses Windows so we need your help for this. Please get in touch if this is something you'd like to help with.