ClickHouse mock data from DDL
ClickHouse DDL often continues after the table name: ON CLUSTER, an engine, PARTITION BY, ORDER BY, and SETTINGS. SchemaSeed skips those clauses and reads the column list.
Types such as Int64, UInt16, Float64, and DateTime map to integers, decimals, and timestamps. Nullable(Int64) stays nullable. LowCardinality(Nullable(String)) is treated as text that may be empty. A PRIMARY KEY written after the column list, or an ORDER BY used as the key, is picked up as the primary key.
INSERT INTO `events` (`id`, `ts`) VALUES (1, '2026-10-14 12:00:00');
ClickHouse is a Pro database. The export uses backticks and writes datetimes with a space.