This reference guide covers all configurable parameters for each SQL tool. The SQL Tool supports five distinct operation types, each designed for specific database interaction patterns.
A general AI-native engine eats specialized tools. InTouch AI does what a dedicated ETL or query-scheduler does — and a config-era tool can never grow an AI core to do what InTouch does. These five SQL operations sit on the deterministic end of the determinism dial: describe the work in your own language to the assistant once, then graduate it to a fully-deterministic, fully-audited SQL tool that runs identical every time at zero AI cost. Every connection below resolves through the encrypted credential vault — referenced by name, never written into a query, never exposed even to the AI itself.
1. Export Tool
Execute SELECT queries and export results directly to files. Use this tool to extract data from any supported database into delimited files for reporting, data feeds, or downstream processing.
| Parameter | Description | Required |
|---|
| Connection | Database connection object (encrypted, reusable) | Yes |
| SQL Query | SELECT statement to execute against the database | Yes |
| Output File | Destination file path for exported data | Yes |
| Delimiter | Field separator: comma, tab, pipe, or custom character | No (default: comma) |
| Include Headers | Include column headers as the first row of output | No (default: yes) |
| Record Limit | Maximum number of rows to export (0 = unlimited) | No |
| File Mode | Overwrite existing file or Append to it | No (default: overwrite) |
| Encoding | Output file character encoding | No (default: UTF-8) |
2. Import Tool
Load data from files into database tables using InTouch AI's proprietary batching technology — up to 1000X faster than standard JDBC inserts. Intelligent column mapping handles schema differences automatically.
| Parameter | Description | Required |
|---|
| Connection | Database connection object | Yes |
| Source File | Input file path containing data to load | Yes |
| Target Table | Destination database table name | Yes |
| Batch Size | Records per batch (proprietary batching for 1000X speed) | No (default: 1000) |
| Column Mapping | Map file columns to table columns (name or position) | No (auto-detected) |
| Header Row | Source file contains a header row | No (default: yes) |
| Delimiter | Field separator in the source file | No (default: comma) |
| Error Threshold | Maximum errors before aborting (0 = abort on any error) | No (default: 0) |
| Truncate First | Clear the target table before loading new data | No (default: no) |
3. SQL Statement Tool
Execute a single SQL statement — INSERT, UPDATE, DELETE, or DDL. Use this for targeted database modifications, trigger resets, inventory updates, or schema changes.
| Parameter | Description | Required |
|---|
| Connection | Database connection object | Yes |
| SQL Statement | SQL to execute (INSERT, UPDATE, DELETE, DDL) | Yes |
| Expected Rows | Expected affected row count for validation (tool fails if mismatch) | No |
4. SQL Stream Tool
Execute a sequence of SQL statements with transaction control. Define statements inline or load them from a file. Ideal for multi-step database operations that must succeed or fail as a unit.
| Parameter | Description | Required |
|---|
| Connection | Database connection object | Yes |
| SQL Source | Inline SQL statements or path to SQL file | Yes |
| Source Type | Inline or File | Yes |
| Stop On Error | Halt execution on first error or continue | No (default: yes) |
| Transaction Mode | Auto-commit each statement or manual commit/rollback at end | No (default: auto) |
5. Stored Procedure Tool
Execute database stored procedures with full parameter support. Works across all supported database types including Oracle, SQL Server, MySQL, PostgreSQL, and DB2.
| Parameter | Description | Required |
|---|
| Connection | Database connection object | Yes |
| Procedure Name | Fully qualified stored procedure name (schema.procedure) | Yes |
| Parameters | List of IN, OUT, and INOUT parameters with data types and values | No |
| Result Set Handling | How to process returned result sets (log, export, or discard) | No |
| Output File | File path for result set output (when exporting results) | No |