sqlclean facts

sqlclean facts parses SQL from standard input and prints the structural facts of each statement as one line of JSON. These are the facts a custom Rule receives, so use it to see what a Rule will read before writing one.

Usage

sqlclean facts [--dialect <DIALECT>] < query.sql

echo "SELECT o.order_id, COUNT(*) AS items FROM orders AS o JOIN order_items AS i ON i.order_id = o.order_id WHERE o.status = 'paid' GROUP BY o.order_id;" \
  | sqlclean facts --dialect duckdb

Flags

Flag	Description
--dialect <DIALECT>	The dialect to parse with (default: generic); see Dialects

facts reads no configuration file and no paths.

Output

The output, pretty-printed here and trimmed to some of the columns:

{
  "statements": [
    {
      "index": 0,
      "kind": "select",
      "start": 0,
      "end": 146,
      "terminated_end": 147,
      "ctes": [],
      "tables": [
        { "start": 42, "end": 48, "name": "orders", "qualified_name": "orders", "alias": "o" },
        { "start": 59, "end": 70, "name": "order_items", "qualified_name": "order_items", "alias": "i" }
      ],
      "joins": [
        { "start": 54, "end": 102, "kind": "INNER", "relation": "order_items", "has_condition": true }
      ],
      "stars": [],
      "functions": [
        { "start": 19, "end": 24, "name": "COUNT" }
      ],
      "columns": [
        { "start": 7, "end": 17, "name": "order_id", "qualifier": "o" },
        { "start": 109, "end": 117, "name": "status", "qualifier": "o" }
      ],
      "literals": [
        { "start": 120, "end": 126, "kind": "string", "text": "'paid'" }
      ]
    }
  ]
}

Offsets count characters from the start of the input, not bytes. Custom Rules receive the same facts with line, column and text added to every node. See Statement facts for every field.

Exit codes

Code	Meaning
0	The facts were printed
2	The SQL does not parse under the dialect, or the dialect is unknown