PostgreSQL schema diagrams from pg_dump output

Point draft at a pg_dump file and read it as a diagram. Schema-qualified names, keys declared through ALTER TABLE, quoted reserved words and multi-word types all parse.

Open the app

What the parser makes of it

• 2 tables
• 11 columns
• 1 relation

• public.users-style schema qualification, kept out of the displayed name
• keys added after the fact by ALTER TABLE … ADD CONSTRAINT, including ON DELETE CASCADE
• reserved words quoted by pg_dump, such as public."order"
• multi-word types — character varying(255), timestamp without time zone
• arrays (text[]) and jsonb defaults containing braces and casts
• SET, OWNER TO, CREATE INDEX and CREATE VIEW, all skipped without complaint

The diagram

The file

This is the project's PostgreSQL test fixture, in full and unedited. Everything above was produced from it.

--
-- PostgreSQL database dump
-- Shaped like pg_dump output: schema-qualified names, keys declared through
-- ALTER TABLE rather than inline, and multi-word types.
--

SET statement_timeout = 0;
SET client_encoding = 'UTF8';

CREATE TABLE public.users (
    id integer NOT NULL,
    email character varying(255) NOT NULL,
    display_name text,
    created_at timestamp without time zone DEFAULT now() NOT NULL,
    settings jsonb DEFAULT '{}'::jsonb,
    tags text[]
);

ALTER TABLE public.users OWNER TO app;

--
-- "order" is a reserved word, so pg_dump quotes it everywhere.
--

CREATE TABLE public."order" (
    id integer NOT NULL,
    user_id integer NOT NULL,
    total numeric(12,2) DEFAULT 0 NOT NULL,
    placed_at timestamp with time zone,
    note text  -- may contain a semicolon; that must not split the statement
);

ALTER TABLE ONLY public.users
    ADD CONSTRAINT users_pkey PRIMARY KEY (id);

ALTER TABLE ONLY public.users
    ADD CONSTRAINT users_email_key UNIQUE (email);

ALTER TABLE ONLY public."order"
    ADD CONSTRAINT order_pkey PRIMARY KEY (id);

ALTER TABLE ONLY public."order"
    ADD CONSTRAINT order_user_id_fkey FOREIGN KEY (user_id) REFERENCES public.users(id) ON DELETE CASCADE;

CREATE INDEX order_user_id_idx ON public."order" USING btree (user_id);

CREATE VIEW public.recent_orders AS
 SELECT o.id, o.total FROM public."order" o WHERE (o.placed_at > now());