Free · No sign-up · Nothing installed

Snowflake to ER Diagram

Paste GET_DDL('SCHEMA', …) output — NUMBER(38,0) identity columns, VARIANT, TIMESTAMP_NTZ and COMMENT = / CLUSTER BY noise handled.

SQL input — CREATE TABLE · DDL · dumpsnowflake

CREATE TABLE analytics.blog.users (
  id NUMBER(38,0) IDENTITY START 1 INCREMENT 1 NOT NULL,
  email VARCHAR(320) NOT NULL,
  display_name VARCHAR(120),
  created_at TIMESTAMP_NTZ(9) NOT NULL DEFAULT CURRENT_TIMESTAMP(),
  PRIMARY KEY (id)
) COMMENT = 'blog user accounts' CLUSTER BY (id);

CREATE TABLE analytics.blog.posts (
  id NUMBER(38,0) IDENTITY START 1 INCREMENT 1 NOT NULL,
  author_id NUMBER(38,0) NOT NULL,
  title VARCHAR(200) NOT NULL,
  body VARCHAR(16777216),
  attrs VARIANT,
  published_at TIMESTAMP_TZ(9),
  PRIMARY KEY (id)
) COMMENT = 'blog posts';

CREATE TABLE analytics.blog.comments (
  id NUMBER(38,0) IDENTITY START 1 INCREMENT 1 NOT NULL,
  post_id NUMBER(38,0) NOT NULL,
  commenter_id NUMBER(38,0) NOT NULL,
  body VARCHAR(16777216) NOT NULL,
  created_at TIMESTAMP_NTZ(9) NOT NULL DEFAULT CURRENT_TIMESTAMP(),
  PRIMARY KEY (id)
) COMMENT = 'post comments';

CREATE TABLE analytics.blog.tags (
  id NUMBER(38,0) IDENTITY START 1 INCREMENT 1 NOT NULL,
  slug VARCHAR(80) NOT NULL,
  PRIMARY KEY (id)
) COMMENT = 'tag dictionary';

CREATE TABLE analytics.blog.post_tags (
  post_id NUMBER(38,0) NOT NULL,
  tag_id NUMBER(38,0) NOT NULL,
  PRIMARY KEY (post_id, tag_id),
  FOREIGN KEY (post_id) REFERENCES analytics.blog.posts (id),
  FOREIGN KEY (tag_id) REFERENCES analytics.blog.tags (id)
) COMMENT = 'post to tag join';

-- the way GET_DDL writes foreign keys:
ALTER TABLE analytics.blog.posts ADD CONSTRAINT fk_posts_author FOREIGN KEY (author_id) REFERENCES analytics.blog.users (id);
ALTER TABLE analytics.blog.comments ADD CONSTRAINT fk_comments_post FOREIGN KEY (post_id) REFERENCES analytics.blog.posts (id);
ALTER TABLE analytics.blog.comments ADD CONSTRAINT fk_comments_commenter FOREIGN KEY (commenter_id) REFERENCES analytics.blog.users (id);

snowflakeSample: · · · · ·

Diagram—

Your diagram renders here.

Parser coverage

What we parse in snowflake

NUMBER(38,0) + IDENTITY — precision kept, auto-increment folded into the PK
Semi-structured types — VARIANT, OBJECT and ARRAY shown verbatim
COMMENT = / CLUSTER BY — stripped as noise, never rendered as columns
Whole schemas — GET_DDL('SCHEMA', 'mydb.analytics') output parsed in one paste

From your database

Get just the schema

SELECT GET_DDL('SCHEMA', 'mydb.analytics');

One statement, every table in the schema. Run it in a worksheet, copy the result, paste it above — nothing ever connects to your account.

FAQ

Questions, answered

How do I generate an ER diagram from Snowflake?

Run SELECT GET_DDL('SCHEMA', 'mydb.my_schema'); in a worksheet and paste the result — every CREATE TABLE comes along in one output.

Does it keep VARIANT and TIMESTAMP_NTZ columns readable?

Yes — VARIANT, OBJECT, ARRAY and NTZ/TZ/LTZ timestamps render the way GET_DDL spells them.

What gets stripped?

COMMENT = '…', CLUSTER BY and other noise clauses — the diagram shows tables, columns, types and keys only.

Does it need a Snowflake connection or key-pair auth?

No. You copy DDL text out of a worksheet; there is no connection to your account, ever.

More converters

Snowflake to ER Diagram

What we parse in snowflake

Get just the schema

Questions, answered

Also available