Query syntax

Supported clauses

Clause	Support	Notes
`SELECT`	Supported	Including `DISTINCT` and `SELECT * EXCLUDE`.
`FROM`	Supported	Tables, views, subqueries, and `LATERAL` subqueries.
`JOIN`	Supported	Inner and outer joins, `CROSS JOIN`, `NATURAL JOIN`, and `JOIN ... USING`.
`WHERE`	Supported	Scalar and `EXISTS` subqueries.
`GROUP BY`	Supported	Including `ALL`, ordinals, `GROUPING SETS`, `ROLLUP`, and `CUBE`.
`HAVING`	Supported
`QUALIFY`	Supported	Filters on window function results.
`ORDER BY`	Supported	Including `NULLS FIRST` and `NULLS LAST`.
`LIMIT` / `FETCH`	Supported	`LIMIT`, `OFFSET`, and `FETCH FIRST ... ROWS ONLY`.
`WITH`	Supported	`WITH RECURSIVE` is not supported.
Window functions	Supported	`PARTITION BY`, explicit frames, and `IGNORE NULLS`.
`LISTAGG ... WITHIN GROUP`	Supported
`TOP n`	Not supported	Use `LIMIT`.
`PIVOT`, `UNPIVOT`	Not supported
`MATCH_RECOGNIZE`	Not supported
`CONNECT BY`	Not supported
`ASOF JOIN`	Not supported
`AT` / `BEFORE` / `CHANGES`	Not supported	Time travel.
`FLATTEN`	Not supported	`LATERAL FLATTEN` and `TABLE(FLATTEN(...))`.
`IDENTIFIER()`	Not supported
`SAMPLE` / `TABLESAMPLE`	Not supported	Row sampling is not available. Use `WHERE RANDOM() < 0.1`.
`SELECT * ILIKE`	Not supported

Fully qualified database.schema.table names are not supported and fail with Database 'mydb' does not exist. Two-part schema.table names work. Most tool-generated Snowflake SQL is fully qualified, so this is usually the first thing a migrated workload hits.
Identifiers are returned in lower case. Snowflake folds unquoted identifiers to upper case, so a column selected as n comes back as N there and n here. A client that looks up result columns by name needs to be case-insensitive.
$$-quoted string literals are not supported.