SQL格式化器
使用清晰缩进格式化 SQL，便于检查子句、连接、筛选条件、注释和表达式。
Make SQL queries easier to review
长 SQL 查询写在一行时很难看清连接、筛选条件和列。本工具提供易读布局，方便在迁移、报表或应用前检查结构。
格式化只改变展示方式，不能证明查询安全、获得授权、性能良好或适合某个数据库版本。请保留原文并检查差异。
在需要时使用包含别名、连接、谓词、聚合或注释的代表性查询。不要因为格式整齐就直接执行未审查语句。
Formatting is not SQL validation
SQL expresses statements, identifiers, literals, operators, clauses, joins, subqueries, and comments. Formatting can place these parts on separate lines so their relationships are easier to read.
A formatter parses enough SQL or token-like text to arrange whitespace. It does not validate permissions, table existence, schema contracts, transaction behavior, query plans, locking, or business meaning.
SQL dialects differ. Keywords, quoting, functions, pagination, casts, procedural syntax, and vendor extensions may be formatted differently or may not be recognized by the selected path. Test with the target database tooling.
How to format SQL
- 把完整且受支持的查询粘贴到输入框。
- 运行格式化，检查 SELECT、JOIN、条件、分组、排序、子查询和注释。
- 与原文比较，确认字面量、标识符、运算符、参数和注释保持正确。
- 执行前使用解析器、代码检查、EXPLAIN、权限审核和测试。
- 示例中移除凭据、个人数据和私有架构细节。
Formatting method
- Read SQL tokens
The tool reads the supplied SQL and identifies keywords, clauses, expressions, delimiters, literals, comments, and nested regions that can be laid out.
- Apply clause layout
Line breaks and indentation make SELECT fields, JOIN conditions, predicates, grouping, and subqueries easier to scan. The whitespace is presentation output.
- Preserve query text
String contents, identifiers, placeholders, operators, comments, and clause order must remain tied to the source. Review a diff when exact text matters.
Exact SQL dialect support, parser recovery, comment handling, identifier quoting, and error wording should be confirmed in a live engine before being described as fixed behavior.
Read the formatted result
A successful result is a readable arrangement of the supplied SQL. Confirm columns, literals, aliases, joins, conditions, comments, and clause order against the source.
A failure means the documented formatting path could not handle the input. Do not infer exact dialect support or parser recovery without runtime evidence.
If a database later rejects or slows the query, investigate syntax, permissions, schema, plan, indexes, and data separately. Formatting is not execution or optimization.
常见错误
Treating indentation as a safety or performance review can hide a broad filter, unintended join, or expensive subquery.
Editing string literals, identifiers, placeholders, or comments while formatting can change application behavior.
Assuming one SQL dialect applies to every database can make formatted vendor syntax misleading or invalid.
Pasting credentials or production data into a sample can expose sensitive information even when the query is read-only.
限制和特殊情况
The documented capability is SQL presentation formatting. It does not guarantee syntax validation, authorization, schema compatibility, performance, transaction safety, or execution results.
Vendor dialects, procedural SQL, quoted identifiers, comments, huge queries, and parser-specific recovery should be tested with the target database tools.
Do not paste confidential SQL or data unless your handling policy permits it. A formatted copy can retain the same secrets and schema details.
Examples
Format a simple SELECT
格式化会改变结果吗？
把完整且受支持的查询粘贴到输入框。
Selected columns and the WHERE clause appear on separate readable lines.
Indent a JOIN
支持哪些方言？
运行格式化，检查 SELECT、JOIN、条件、分组、排序、子查询和注释。
The table aliases and join condition remain visible in the formatted query.
Keep a string literal
工具会验证查询吗？
与原文比较，确认字面量、标识符、运算符、参数和注释保持正确。
The spaces inside the string literal remain part of the value.
Indent aggregation
可以格式化子查询吗？
执行前使用解析器、代码检查、EXPLAIN、权限审核和测试。
The aggregate, grouping, and ordering clauses become distinct review points.
Keep a comment
注释会保留吗？
示例中移除凭据、个人数据和私有架构细节。
The comment remains associated with the statement in the formatted view.
Review a placeholder
参数如何处理？
把完整且受支持的查询粘贴到输入框。
The placeholder stays in the predicate for the calling application to bind.
Frequently asked questions
格式化会改变结果吗？
格式化应只改变展示；执行前请用 diff 确认文本。
支持哪些方言？
支持范围取决于解析器和目标数据库。
工具会验证查询吗？
不会。语法、权限、架构、计划和测试需要分别检查。
可以格式化子查询吗？
如果结构可识别可以；请检查括号和方言。
注释会保留吗？
目标是保留注释，但复杂注释需要测试。
参数如何处理？
参数应保持不变，不要替换成真实值。
可以直接执行输出吗？
不要，先在受控环境验证。
如何比较前后文本？
保留原文，格式化副本并逐行查看 diff。