# sql

Connect to a SQL database, execute a query, and emit each result row as an event.  When not wired to any inputs, the query executes once on pipeline start (with `{START: true}` as the event).  When wired to an input, the query executes once per received event.

Supported drivers: `postgres`, `mysql`.

## Config

**Required:**
- `driver`: Database driver — `postgres` or `mysql`
- `connection`: Connection URL string (e.g. `postgres://user:pass@host/db`), or a JS function `() => string`
- `query`: JS function `(event) => string` — returns the SQL string to execute

**Optional:**
`SQL(string)` is available as a no-op identity helper inside the query function for compatibility with tagged-template patterns.

## Examples

Static query on startup:

```yaml
pipeline:
  sql:
    config:
      driver: postgres
      connection: "() => 'postgres://' + process.env.DB_URL"
      query: "(event) => 'SELECT * FROM tasks WHERE status = \\'queued\\''"
```

Query per input event (e.g. from a scheduler):

```yaml
pipeline:
  scheduler:
    config:
      cron: '*/5 * * * *'
  sql:
    inputs: [scheduler]
    config:
      driver: postgres
      connection: postgres://localhost/mydb
      query: "function(event, db, SQL) { return SQL('SELECT * FROM logs WHERE ts > now() - interval \\'5 minutes\\'') }"
```

`process.env` is available inside JS functions for reading environment variables.