AI Chat

Dory AI Chat is designed for database work. It can use connection, schema, table, SQL, and result context to help you move from a question to executable SQL and follow-up analysis.

Why Use Dory AI Chat

• It understands database context instead of answering as a generic chatbot.
• It can generate SQL from a natural-language request.
• It can explain existing SQL and query results.
• It can help fix errors and suggest safer or faster queries.
• It works alongside Explorer and SQL Console.

What Context AI Chat Understands

• Selected database connection.
• Visible schemas, tables, and columns.
• Current SQL editor content.
• Recent query errors or result shape.
• Saved queries or workflow context when available.

Core Capabilities

1. Generate SQL

Describe the business question, tables, filters, time range, and output columns. Dory can draft SQL that you review in SQL Console.

2. Explain SQL

Paste SQL or use the current editor content and ask AI to explain joins, filters, aggregations, and potential risks.

3. Fix SQL Errors

When a query fails, ask AI to explain the error and propose a corrected version. Always compare the fix with Explorer schema.

4. Optimize SQL

Ask AI to reduce scanned data, simplify joins, add safer filters, or rewrite aggregations.

5. Continue Analysis

After reviewing results, ask follow-up questions such as grouping by another dimension, comparing time periods, or producing chart-ready output.

Recommended Usage

Use with SQL Console (Recommended)

Use AI Chat to draft or improve SQL, then run and validate it in SQL Console.

Ask Standalone Questions

Use AI Chat to understand table purpose, column meaning, or possible analysis paths.

Use Table Context

Start from Explorer when possible so AI has the table and column names you intend to use.

High-Quality Prompt Examples

• Generate SQL to count daily active users for the last 30 days. Return date and user count.
• Explain this SQL and point out any risky filters.
• Fix this ClickHouse error and keep the result columns unchanged.
• Rewrite this query so the result is suitable for a line chart by day.

Best Practices

• Give exact table names and output expectations.
• Review every generated query before execution.
• Start with read-only or limited queries.
• Use Explorer to verify table and column names.
• Ask AI to preserve semantics when optimizing SQL.

Recommended Next Step

Continue with Generate SQL with Dory AI.