9.0.0-beta.6

Breaking changes

• LIMIT & OFFSET keywords are no more supported in all dialects #362

Bugfixes

• Fix missing type-definition file when importing sql-formatter in TypeScript #353

Enhancements

• Allow # as first char for Redshift identifiers #358
• Allow MySQL and MariaDB identifiers to begin with a number #339
• Improved formatting of SELECT DISTINCT #362
• Added support for several variations in SELECT query clauses (like WITH RECURSIVE in addition to plain WITH) #362
• General overhaul of query syntax support #362

9.0.0-beta.5

Bugfixes:

• Fix multiword commands formatting with keywordCase:upper #356

Removals:

• Removed support for lots of questionable DB2 keywords #352

Internals:

• Rename BINARY_COMMAND to SET_OPERATION #351

9.0.0-beta4

Bugfixes:

• Fix BETWEEN..AND formatting in tabular style #350
• Fix double-quoted strings support for Spark #331

Enhancements:

• Improved formatting of PL/SQL tbl@dblink syntax #338
• Added support for PL/SQL Q strings #342
• Improved Trino row pattern formatting #333
• Revised reserved keyword lists in multiple dialects #348
• Revised supported joins and set operations in multiple dialects #346
• Revised supported string types in multiple dialects #332

Internals:

• Utility for auto-generation of syntax variants #327

9.0.0-beta3

New SQL dialect: Trino

• Thanks for the work of Boris Verkhovskiy #325 #297

9.0.0-beta2

Improvements

• Improved support for various BigQuery DDL statements (thanks to Ahmad Khan #319)

Bugfixes

• Fixed aliasAs option inserting AS keywords to BigQuery functions #309
• Fixed parameter substitution not working in BETWEEN expressions #318

Internals

• Placed each language.formatter.ts file to separate directory and separated functions and keywords to separate files #315
• Removed RESERVED_JOIN_CONDITION token type.

9.0.0-beta1

New approach to function call matching

This release brings a major change to how function calls are detected:

• Only known builtin function names are now formatted like TRIM()
• Any other function calls (e.g. user-defined functions) get formatted as MY_FUNC ()

This fixes a long-standing problem where the format of input SQL effected output SQL. For example, if input contained TRIM (), then it was formatted as TRIM () and when it contained TRIM() then it was formatted as TRIM(). This in turn lead to problems where re-formatting an already formatted SQL could lead to different result. #140

8.2.0

New feature

• Added support for formatting WINDOW clause in all dialects supporting it #305

8.1.0

Bugfixes

• Fix formatting of DESC inside ORDER BY for MariaDB #298
• Fix multiple problems with LIMIT clause formatting #301 #303

Improvements

• No more adding spaces around :: operator.
• Add support for multiple PostgreSQL JSON and other operators #296
• Add support for N1QL operator: ||.
• Add support for DB2 operators: ¬=, ¬>, ¬<.
• Add support for SQLite operators: ->, ->>.
• Add support for ~, #, ?, ! operators in several dialects.

Internals

• Reorganize code in core/ dir into lexer/, parser/, formatter/ directories.
• Major refactoring of Tokenizer.
• Several new token types: DELIMITER, COMMA, IDENTIFIER, QUOTED_IDENTIFIER, NAMED_PARAMETER, QUOTED_PARAMETER, INDEXED_PARAMETER, POSITIONAL_PARAMETER.

8.0.2

Bugfixes

• Fix BigQuery STRUCT formatting #279
• Fix BigQuery FROM clause operators formatting #279

8.0.1

Bugfix

• Support arrays in PostgreSQL #275