SHOW SEMANTIC FACTS¶

Lists facts (named row-level expressions) registered in one or all semantic views. Each row describes a single fact with its name, source table, data type, synonyms, and comment. Views that have no facts defined return no rows.

Syntax¶

SHOW SEMANTIC FACTS
    [ LIKE '<pattern>' ]
    [ IN { <name> | ACCOUNT
         | DATABASE [ <database_name> ]
         | SCHEMA [ [<database_name>.]<schema_name> ] } ]
    [ STARTS WITH '<prefix>' ]
    [ LIMIT <rows> ]

All clauses are optional. When multiple clauses appear, they must follow the order shown above.

Statement Variants¶

SHOW SEMANTIC FACTS: Returns facts across all registered semantic views, sorted by semantic view name and then fact name. Views with no FACTS clause are omitted from the result.
SHOW SEMANTIC FACTS IN <name>: Returns facts for the specified semantic view only, sorted by fact name. Returns an error if the view does not exist. Returns an empty result if the view has no facts.

Parameters¶

<name>: The name of the semantic view. Required only for the single-view form (IN clause). Returns an error if the view does not exist.

Optional Filtering Clauses¶

LIKE '<pattern>'

Filters facts to those whose name matches the pattern. Uses SQL LIKE pattern syntax: % matches any sequence of characters, _ matches a single character. Matching is case-insensitive (the extension maps LIKE to DuckDB’s ILIKE). The pattern must be enclosed in single quotes.

IN ...

Scopes the listing. The alternatives are mutually exclusive – IN appears at most once, so a view name and a schema cannot both be given.

IN <name>: Returns facts for that semantic view only.
IN ACCOUNT: Returns everything, the same as omitting IN entirely. Accepted for Snowflake compatibility; DuckDB has no account.
IN DATABASE [ <database_name> ]: Filters facts to semantic views in that database, or the current database when the name is omitted.
IN SCHEMA [ [<database_name>.]<schema_name> ]: Filters facts to semantic views in that schema, or the current schema when the name is omitted. Qualifying the schema with a database matches on both, so a same-named schema in another database is excluded.

Schema and database names follow DuckDB’s identifier rule: quotes are stripped and case is ignored. A view named schema or database must be quoted (IN "schema") to be read as a view name rather than as the scope keyword.

STARTS WITH '<prefix>'

Filters facts to those whose name begins with the prefix. Matching is case-sensitive. The prefix must be enclosed in single quotes.

LIMIT <rows>

Restricts the output to the first rows results. Must be a non-negative integer; LIMIT 0 is accepted and returns no rows.

When LIKE and STARTS WITH are both present, a fact must satisfy both conditions (they are combined with AND).

Warning

Clause order is enforced. LIKE must come before IN, and STARTS WITH must come after IN. Placing clauses out of order produces a syntax error.

Output Columns¶

Returns one row per fact with 8 columns:

Column	Type	Description
`database_name`	VARCHAR	The DuckDB database containing the semantic view.
`schema_name`	VARCHAR	The DuckDB schema containing the semantic view.
`semantic_view_name`	VARCHAR	The semantic view this fact belongs to.
`table_name`	VARCHAR	The physical table name the fact is scoped to.
`name`	VARCHAR	The fact name as declared in the `FACTS` clause.
`data_type`	VARCHAR	The declared output type. Empty for every view created since v0.10.0 – no surface can declare a type and nothing infers one. Populated only for views stored before that release. See Reported Data Types.
`synonyms`	VARCHAR	JSON array of synonym strings (e.g., `["discounted_price"]`). Empty string if no synonyms are set.
`comment`	VARCHAR	The fact comment text. Empty string if no comment is set.

Examples¶

List facts for a single view:

Given a semantic view orders_sv with one fact:

SHOW SEMANTIC FACTS IN orders_sv;

┌───────────────┬─────────────┬──────────────────────┬────────────┬────────────┬────────────────┬──────────┬─────────┐
│ database_name │ schema_name │ semantic_view_name   │ table_name │ name       │ data_type      │ synonyms │ comment │
├───────────────┼─────────────┼──────────────────────┼────────────┼────────────┼────────────────┼──────────┼─────────┤
│ memory        │ main        │ orders_sv            │ orders     │ raw_amount │                │          │         │
└───────────────┴─────────────┴──────────────────────┴────────────┴────────────┴────────────────┴──────────┴─────────┘

data_type is empty here because no surface can declare a member’s output type: the SQL DDL has no clause for it, and the YAML output_type field was withdrawn because GET_DDL could not carry it (a restored view silently lost the cast). Nothing infers one either – v0.10.0 removed the define-time inference pass – so the column is populated only for views stored before that release. Reporting the type an expression actually produces would require probing it on the read side, at SHOW bind time; that is a known limitation and is not implemented today. See Reported Data Types.

List facts across all views:

SHOW SEMANTIC FACTS;

Views with no facts are omitted.

Filter by pattern with LIKE (case-insensitive):

Find all facts whose name contains “amount”:

SHOW SEMANTIC FACTS LIKE '%amount%';

Because LIKE is case-insensitive, LIKE '%AMOUNT%' produces the same results.

Filter by schema:

SHOW SEMANTIC FACTS IN SCHEMA main;

Chained facts:

Facts can reference other facts. Consider a view with two chained facts:

CREATE SEMANTIC VIEW tpch_analysis AS
TABLES (
    li AS line_items PRIMARY KEY (id)
)
FACTS (
    li.net_price  AS li.extended_price * (1 - li.discount),
    li.tax_amount AS li.net_price * li.tax_rate
)
DIMENSIONS (
    li.status AS li.status
)
METRICS (
    li.revenue AS SUM(li.net_price)
);

SHOW SEMANTIC FACTS IN tpch_analysis;

┌───────────────┬─────────────┬──────────────────────┬────────────┬────────────┬────────────────┬──────────┬─────────┐
│ database_name │ schema_name │ semantic_view_name   │ table_name │ name       │ data_type      │ synonyms │ comment │
├───────────────┼─────────────┼──────────────────────┼────────────┼────────────┼────────────────┼──────────┼─────────┤
│ memory        │ main        │ tpch_analysis        │ line_items │ net_price  │                │          │         │
│ memory        │ main        │ tpch_analysis        │ line_items │ tax_amount │                │          │         │
└───────────────┴─────────────┴──────────────────────┴────────────┴────────────┴────────────────┴──────────┴─────────┘

Both facts show an empty data_type, including net_price, whose expression uses only physical columns: no type is inferred for either. Chained references (tax_amount references li.net_price) are resolved at query expansion time regardless.

Error: view does not exist:

SHOW SEMANTIC FACTS IN nonexistent_view;

Error: semantic view 'nonexistent_view' does not exist