Sample Data

Opteryx includes one built-in relation for demonstration and testing:

$planets (20 columns, 9 rows)

SELECT *
  FROM $planets;

$planets is temporal — using Time Travel returns different results depending on the date. Uranus was discovered in 1781 and Pluto in 1930, so querying before those dates returns fewer rows:

sqlOpen in Studio

SELECT name
  FROM $planets
   TIMESTAMP AS OF '1700-01-01'::TIMESTAMP;

Other internal relations exist prefixed with $ (such as $variables and $user). These are not intended for end-user queries — their structure and availability are not guaranteed.

Starting From a Ready-Made Dataset

The $ relations are tiny — enough to try an expression, not enough to write a realistic query against. For that, fork a dataset that already exists with CREATE TABLE ... CLONE:

sql

CREATE TABLE personal.alice.lineitem CLONE samples.tpch_sf1.lineitem;

Nothing is copied: the new dataset borrows the original's files, so the statement costs the same whether the source holds a megabyte or a terabyte. Unlike the $ relations, what you get is an ordinary dataset — you can write to it, time-travel it and drop it, and it occupies no storage of your own until you change it.

Use ALTER TABLE ... RESYNC to pick up changes the source has made since, and ALTER TABLE ... DETACH to take a private copy and end the relationship.