duckdb.query_file

Run a SQL file over the connected tables with DuckDB, without network. The tables are views named t0, t1, ... in connection order. The file may be several statements; the result of the last is the output. Runs queries written by ai.sql, saved from an earlier run, or uploaded.

Kind: standard
Network: no
Memory: 8Gi

Inputs🔗

Name	Type	Required	Default	Description
`query`	file (SQL)	yes		The SQL to run, in DuckDB's dialect.
`tables`	collection of file (Parquet)	no		Tables to query (Parquet or CSV), bound as t0, t1, ... in order. Leave unconnected for a query that makes its own data.
`output_format`	string	no	`parquet`	File format of the result. One of `parquet`, `csv`.

Outputs🔗

Name	Type	Description
`result`	file (Parquet)	The result of the last statement.

Example🔗

- id: "@query_file"
  name: duckdb.query_file
  inputs: [{block: "@write_query", output: query, as: query}, {block: "@pivot", output: result, as: tables}]
  args: {}