SQL Quick Reference

Dialect support, clause order, JOIN types, operators, best practices and the validation checklist to run before returning any SQL.

#Supported Database Dialects

#Query Structure Reference

#SELECT Statement Components

SELECT [DISTINCT] columns
FROM table
[JOIN other_table ON condition]
[WHERE filter_condition]
[GROUP BY grouping_columns]
[HAVING aggregate_condition]
[ORDER BY sort_columns]
[LIMIT n OFFSET m]

#Common Clause Order

1. SELECT - What columns to return
2. FROM - Source table(s)
3. JOIN - Additional tables with relationships
4. WHERE - Row-level filtering (before grouping)
5. GROUP BY - Aggregate grouping
6. HAVING - Aggregate filtering (after grouping)
7. ORDER BY - Result sorting
8. LIMIT/OFFSET - Pagination

#Best Practices Summary

#Query Writing

• Use explicit column names, avoid SELECT *
• Alias tables and complex expressions
• Use parameterized queries to prevent SQL injection
• Format queries for readability

#Performance

• Index columns used in WHERE, JOIN, ORDER BY
• Avoid functions on indexed columns in WHERE
• Use EXISTS instead of IN for subqueries
• Limit result sets with appropriate filtering

#Maintainability

• Use CTEs for complex multi-step queries
• Add meaningful aliases
• Break complex logic into views or functions
• Document business logic in comments

#Error Handling Pattern

-- PostgreSQL
BEGIN;
    -- your queries here
    IF condition THEN
        RAISE EXCEPTION 'Error message';
    END IF;
COMMIT;

-- SQL Server
BEGIN TRY
    BEGIN TRANSACTION;
    -- your queries here
    COMMIT TRANSACTION;
END TRY
BEGIN CATCH
    ROLLBACK TRANSACTION;
    THROW;
END CATCH;

#Quick Reference Cards

#Aggregation Functions

#JOIN Types

#Common Operators

#Validation Checklist

Before outputting any SQL:

• Syntax is valid for target dialect
• Table and column names exist
• JOIN conditions are complete (no cartesian products)
• WHERE clause has appropriate filters
• GROUP BY includes all non-aggregated columns
• ORDER BY columns are in SELECT or valid
• Parameters are properly escaped/parameterized
• Query handles NULL values appropriately