# PostgreSQL

This output supports efficient batch loading of data into PostgreSQL tables.

## Prerequisites

The PostgreSQL Output requires:
- An existing PostgreSQL database
- A database user with permissions to write to tables
- The target table must exist with appropriate schema

## Configuration

The PostgreSQL Output can be configured using either individual connection parameters or a connection string.

### Settings

| Setting | Type | Required | Default | Description |
|---------|------|----------|---------|-------------|
| host | string | No* | - | The host of the PostgreSQL database |
| port | integer | No* | - | The port of the PostgreSQL database |
| user | string | No* | - | The user to connect to the PostgreSQL database |
| database | string | No* | - | The database name to connect to |
| table | string | Yes | - | The table name to write data to |
| column_names | array[string] | No | - | The column names to write data to, must match the root fields of the data. If not provided all root fields will be used |

*Required if connection_string is not provided

### Secrets

| Setting | Type | Required | Description |
|---------|------|----------|-------------|
| connection_string | text | No* | The connection string to connect to the PostgreSQL database. This will be used over other fields if provided |
| password | text | No* | The password for the PostgreSQL user |

*Either connection_string or individual connection parameters (including password) must be provided

## Data Loading

The PostgreSQL Output uses efficient batch loading with the following characteristics:

1. **Batch Processing**
   - Records are automatically batched for efficient loading
   - Default batch size: 100 records
   - Maximum batch data size: 1 MiB
   - Batch processing interval: 5 seconds

2. **Column Handling**
   - Automatically maps JSON fields to table columns
   - Supports explicit column mapping via column_names setting
   - Handles missing fields by inserting NULL values
   - Uses the first record's schema if no column names are specified

## Best Practices

### Data Types

- Ensure PostgreSQL column types match your data
- Consider using JSONB for complex nested structures
- Use appropriate numeric types for precision requirements