Go FDW for PostgreSQL

An experimental Go project template for building PostgreSQL Foreign Data Wrappers (FDW).

Tested with PostgreSQL v9.6 and Go 1.8.1 on Ubuntu x64.

Supports:

• Table scan
• EXPLAIN
• Table options

Contributions are welcome!

Getting started

Module entry point is defined in fdw.go file (see SetTable). This file contains a basic working example, so give it a try. Later you will need to rewrite it to suit your needs.

Build

The easiest way to build Postgres extension is to run dennwc/go_fdw Docker image:

docker run --rm -v $GOPATH:/gopath -v $PWD/fdw.go:/build/fdw.go -v $PWD/out:/out dennwc/go_fdw

This command will mount your ./fdw.go and GOPATH to the container, build an extension and copy it to ./out folder.

If you don't use Docker, check Dockerfile to see what commands are needed to manually setup your build environment.

Install

To install an extension just copy it to your Postgres installation. At the end of the build process you'll see following lines:

/usr/bin/install -c -m 755  go_fdw.so '/usr/lib/postgresql/9.6/lib/go_fdw.so'
/usr/bin/install -c -m 644 .//go_fdw.control '/usr/share/postgresql/9.6/extension/'
/usr/bin/install -c -m 644 .//go_fdw--1.0.sql  '/usr/share/postgresql/9.6/extension/'

You can execute the same commands to install an extension locally.

Testing

Execute following SQL statements to load an extension:

-- should match an extension lib name
-- creates a "go-fdw" FDW server record automatically (see go_fdw--1.0.sql)
CREATE EXTENSION go_fdw;

-- create a foreign table for an extension
CREATE FOREIGN TABLE public.gotest (
  id INTEGER NOT NULL,
  name text NOT NULL
)
SERVER "go-fdw"
OPTIONS (foo 'bar');

And finally, run the query:

SELECT * FROM gotest;

Note: After extension is loaded, you'll need to restart Postgres to test any changes to the code. Database will not reload shared library file automatically.

Hacking

Project is under development and lacks few features, so feel free to send a PR or file an issue :)

An official documentation for FDW callbacks can be found here. And the documentation for Postgres sources is here.