PolarDB:NL2SQL_SUMMARY: Result summarization

Workflow

The core process of NL2SQL_SUMMARY involves two stages. First, it uses natural language to SQL (NL2SQL) model to convert natural language into an SQL query statement. Then, it uses the parsed SQL statement to generate an analytical summary.

1. Generate an SQL query statement from natural language using the NL2SQL model.

   /*polar4ai*/SELECT * FROM PREDICT (MODEL _polar4ai_nl2sql, select 'Query the credit score of each merchant') WITH (basic_index_name='schema_index',pattern_index_name='pattern_index');

   The generated SQL statement is as follows. Copy the SQL statement generated.

   SELECT merchname AS merchant_name, limitscore AS credit_score FROM hkrt_merchant_info ORDER BY limitscore;

2. Use the NL2SQL_SUMMARY model to summarize and analyze the SQL query results. Replace <SQL_statement> with the code copied from Step 1.

   /*polar4ai*/SELECT * FROM PREDICT (MODEL _polar4ai_nl2sql_summary,<SQL_statement> WITH (usr_query = <usr_query>);

   Parameters

   Parameter	Description	Example
   <usr_query>	The user's question. This parameter clarifies the requirements for data analytics and summarization.	'Merchant type distribution'.
   <SQL_statement>	The SQL query statement generated by the NL2SQL model. It is used to fetch the data for summarization. Do not add a semicolon at the end.	SELECT merchname AS merchant_name, limitscore AS credit_score FROM hkrt_merchant_info ORDER BY limitscore.

   Example

   /*polar4ai*/SELECT * FROM PREDICT (MODEL _polar4ai_nl2sql_summary,
   SELECT merchname AS merchant_name, limitscore AS credit_score FROM hkrt_merchant_info ORDER BY limitscore
    ) WITH (usr_query='Query the credit score of each merchant');

   Output:

   This query returned the credit scores for 15 merchants, ranging from 60 to 95. The lowest score is 60 (Future Star Merchant), and the highest is 95 (Tech City Merchant). Overall, most merchants have scores above 70, indicating good credit standing.

   SQL query results:

   No.	Credit Score	Merchant Name
   1	60	Future Star Merchant
   2	65	Stellar Tech Merchant
   3	68	Innovative Future Merchant
   4	70	Stellar Glory Merchant
   5	72	Stellar Dream Merchant
   6	75	Future Tech Merchant
   7	80	Dream Tech Merchant
   8	82	Dream Tech Merchant
   9	85	Xiaotian Tech Merchant
   10	85	Glorious Dream Merchant
   11	88	Glorious Future Merchant
   12	90	Glorious Tech Merchant
   13	90	Glorious Star Merchant
   14	92	Tech Star Merchant
   15	95	Tech City Merchant

Best practices

• You can combine NL2SQL and NL2SQL_SUMMARY to automate the entire flow from natural language queries to intelligent summaries.

• Ensure that your questions are clear. Providing a clear usr_query improves the quality and relevance of the summary.

• For better visualization, use the NL2Chart feature to generate charts that complement the summary.

• Recommended scenarios:

  Scenario	Typical application
  Data overview	Quickly understand the data distribution across different categories or time periods in a database.
  Decision support	Provide data-driven summary recommendations for business decisions.
  Anomaly detection	Detect abnormal categories or trends and suggest possible causes.
  Report writing	Automatically generate key sections of data reports to improve efficiency.

NL2SQL_SUMMARY quickly extracts key information from raw data and creates structured summaries. This provides a solid foundation for subsequent data analytics and decision-making.