Skip to content

Review and Explain SQL

Who this is for

For users who have generated SQL and now need quality evaluation, risk inspection, and team communication support.

What this solves

You can use AI Review to identify structural risk first, then use Explain selection to quickly clarify key statement intent and reduce review communication cost.

Prerequisites

  • Valid DDL output already exists on the right side.
  • You want quality review or segment explanation for current SQL.

Steps

  1. In the Table DDL area, click AI Review. Result: the system analyzes current DDL and generates scores and suggestions.
  2. Inspect issue descriptions and suggestions in the review result. Result: you can quickly identify high-risk points and improvements.
  3. Click Apply on executable suggestions. Result: suggestions are written back into field or index configuration and become ready for further verification.
  4. Click the history icon in the review area. Result: you can view historical review records for the same table and revisit suggestions.
  5. Select one SQL snippet in the code area. Result: the Explain selection entry appears.
  6. Click Explain selection to view AI explanation. Result: you can quickly understand snippet intent and use it in review communication.
  7. When you need to supplement comments for tables or fields, request AI to generate Chinese business comments. Result: AI automatically generates Chinese comments for tables and fields and fills them into the corresponding positions.

Done when

  • Current DDL has completed at least one review round with readable conclusions.
  • Key suggestions are either applied as needed or explicitly marked as not adopted.
  • Team members can understand key SQL snippets through explanations.

Common pitfalls and failure handling

  • No valid DDL means review cannot run: complete table structure and generate DDL first.
  • "Applicable" does not mean "must apply": suggestions involving business rules require manual confirmation.
  • No response from Explain selection: confirm the selected content is actually SQL text.
  • Empty review history: usually this table has not completed a successful review yet.