sqlglot 30.25.0__tar.gz → 30.26.0__tar.gz

{sqlglot-30.25.0 → sqlglot-30.26.0}/CHANGELOG.md

@@ -1,6 +1,113 @@
 Changelog
 =========
 
+## [v30.25.0] - 2026-10-26
+### :boom: BREAKING CHANGES
+- due to [`c90a322`](https://github.com/tobymao/sqlglot/commit/c90a322c329d2b1003bf73b848f0d9e17f1a3d67) - support non-identifier column parts in compiled code *(PR [#8710](https://github.com/tobymao/sqlglot/pull/8710) by [@georgesittas](https://github.com/georgesittas))*:
+
+  support non-identifier column parts in compiled code (#8710)
+
+- due to [`78b24a7`](https://github.com/tobymao/sqlglot/commit/78b24a744adb8d5dc0eb285f067087338b60f266) - transpile hex SHA2 hashes with ENCODE(..., 'hex') and parse SHA256 as a digest [CLAUDE] *(PR [#8764](https://github.com/tobymao/sqlglot/pull/8764) by [@vidigoat](https://github.com/vidigoat))*:
+
+  transpile hex SHA2 hashes with ENCODE(..., 'hex') and parse SHA256 as a digest [CLAUDE] (#8764)
+
+- due to [`0e7b2da`](https://github.com/tobymao/sqlglot/commit/0e7b2da5d199388e84908aef13c53e535f1cf715) - preserve grouping when removing redundant casts *(PR [#8787](https://github.com/tobymao/sqlglot/pull/8787) by [@georgesittas](https://github.com/georgesittas))*:
+
+  preserve grouping when removing redundant casts (#8787)
+
+- due to [`f0b4347`](https://github.com/tobymao/sqlglot/commit/f0b4347b84b71bfd406e2fe831656a6d03351110) - coerce `VALUES` column types across all rows *(PR [#8775](https://github.com/tobymao/sqlglot/pull/8775) by [@drzippie](https://github.com/drzippie))*:
+
+  coerce `VALUES` column types across all rows (#8775)
+
+- due to [`9d2966e`](https://github.com/tobymao/sqlglot/commit/9d2966ed3f8a7f0d370846dbae1579b40ddc2924) - defer scope annotation until qualification is complete *(PR [#8696](https://github.com/tobymao/sqlglot/pull/8696) by [@georgesittas](https://github.com/georgesittas))*:
+
+  defer scope annotation until qualification is complete (#8696)
+
+
+### :sparkles: New Features
+- [`a220ffb`](https://github.com/tobymao/sqlglot/commit/a220ffb98a2f0ba290ff09c75154a73f09ff57b1) - **duckdb**: support NEAREST joins *(PR [#8722](https://github.com/tobymao/sqlglot/pull/8722) by [@georgesittas](https://github.com/georgesittas))*
+- [`5e0f716`](https://github.com/tobymao/sqlglot/commit/5e0f7169307a4c77fe4cd1eba484eed47d124c4e) - **optimizer**: back-project columns to single unknown-schema TVF *(PR [#8677](https://github.com/tobymao/sqlglot/pull/8677) by [@fivetran-kwoodbeck](https://github.com/fivetran-kwoodbeck))*
+- [`43fbff2`](https://github.com/tobymao/sqlglot/commit/43fbff2037b3c4aa4459ca5bce111c6a1c0c1559) - **optimizer**: Fix positional renaming to have proper type *(PR [#8725](https://github.com/tobymao/sqlglot/pull/8725) by [@fivetran-amrutabhimsenayachit](https://github.com/fivetran-amrutabhimsenayachit))*
+- [`1852fd3`](https://github.com/tobymao/sqlglot/commit/1852fd31961ac51afc38c923e33af84a447adced) - **starrocks**: parse blacklist and TRANSLATE TRINO statements as commands [CLAUDE] *(PR [#8732](https://github.com/tobymao/sqlglot/pull/8732) by [@nytai](https://github.com/nytai))*
+- [`e3b3e23`](https://github.com/tobymao/sqlglot/commit/e3b3e23c644728be22e718b6ca838f753757592f) - **duckdb**: normalize explicit join types *(PR [#8759](https://github.com/tobymao/sqlglot/pull/8759) by [@georgesittas](https://github.com/georgesittas))*
+- [`46af6e4`](https://github.com/tobymao/sqlglot/commit/46af6e468564c271f7cc748dda68a6d813a683f6) - **duckdb**: v2 [dis]connect statement *(commit by [@georgesittas](https://github.com/georgesittas))*
+- [`c514be0`](https://github.com/tobymao/sqlglot/commit/c514be049b382191d6ba010c2af4765d9b8cca57) - **duckdb**: parse SQL trigger bodies *(PR [#8796](https://github.com/tobymao/sqlglot/pull/8796) by [@georgesittas](https://github.com/georgesittas))*
+
+### :bug: Bug Fixes
+- [`67961af`](https://github.com/tobymao/sqlglot/commit/67961afbed7c986dfe26ab82db89a25814c1bdcb) - **optimizer**: preserve unresolved JSON paths during alias matching *(PR [#8693](https://github.com/tobymao/sqlglot/pull/8693) by [@georgesittas](https://github.com/georgesittas))*
+- [`c90a322`](https://github.com/tobymao/sqlglot/commit/c90a322c329d2b1003bf73b848f0d9e17f1a3d67) - support non-identifier column parts in compiled code *(PR [#8710](https://github.com/tobymao/sqlglot/pull/8710) by [@georgesittas](https://github.com/georgesittas))*
+- [`d335f89`](https://github.com/tobymao/sqlglot/commit/d335f8985e59ad986408604ce92733e6834f24eb) - **generator**: parenthesize CONCAT paths in JSON arrow operators [CLAUDE] *(PR [#8701](https://github.com/tobymao/sqlglot/pull/8701) by [@vidigoat](https://github.com/vidigoat))*
+- [`e6c7d52`](https://github.com/tobymao/sqlglot/commit/e6c7d5275707635d7f5cbf08ba6f8095126ab669) - **parser**: parse the column list after an INSERT target alias as the INSERT's [CLAUDE] *(PR [#8712](https://github.com/tobymao/sqlglot/pull/8712) by [@vidigoat](https://github.com/vidigoat))*
+- [`83479d0`](https://github.com/tobymao/sqlglot/commit/83479d061b1b92836cc41b6fdf266f575a72dc3f) - duckdb v2 compat *(PR [#8717](https://github.com/tobymao/sqlglot/pull/8717) by [@georgesittas](https://github.com/georgesittas))*
+- [`092e29a`](https://github.com/tobymao/sqlglot/commit/092e29a82fdb153c6c1925627b0b4a7955398c9b) - **databricks**: preserve timestamp precision during SQL generation *(commit by [@georgesittas](https://github.com/georgesittas))*
+- [`2b57c00`](https://github.com/tobymao/sqlglot/commit/2b57c00a114ad74f8d016f499b0cc97c7ffca9d5) - **sqlite**: parse multi-argument MAX/MIN as GREATEST/LEAST *(PR [#8734](https://github.com/tobymao/sqlglot/pull/8734) by [@vidigoat](https://github.com/vidigoat))*
+- [`83c09cd`](https://github.com/tobymao/sqlglot/commit/83c09cdf3cdd42a6a28a9e7514fdef0a9e202d77) - **oracle**: allow non-reserved SAMPLE and MATCH_RECOGNIZE as identifiers [CLAUDE] *(PR [#8733](https://github.com/tobymao/sqlglot/pull/8733) by [@adrianlzt](https://github.com/adrianlzt))*
+  - :arrow_lower_right: *fixes issue [#8715](https://github.com/tobymao/sqlglot/issues/8715) opened by [@adrianlzt](https://github.com/adrianlzt)*
+- [`069a751`](https://github.com/tobymao/sqlglot/commit/069a75175b3914ccaed99162eaae9b817482da48) - **clickhouse**: natural join [#8739](https://github.com/tobymao/sqlglot/pull/8739) *(commit by [@georgesittas](https://github.com/georgesittas))*
+- [`a97d035`](https://github.com/tobymao/sqlglot/commit/a97d03579f530531451e773823d823404cb6bedd) - **duckdb**: keep the - end placeholder of a slice with a step [CLAUDE] *(PR [#8736](https://github.com/tobymao/sqlglot/pull/8736) by [@vidigoat](https://github.com/vidigoat))*
+- [`30ceae5`](https://github.com/tobymao/sqlglot/commit/30ceae53a0049859d9fd36b2ec1336a9cad7d4d2) - **optimizer**: preserve semantics when unnesting subqueries in JOIN ON *(PR [#8742](https://github.com/tobymao/sqlglot/pull/8742) by [@georgesittas](https://github.com/georgesittas))*
+  - :arrow_lower_right: *fixes issue [#8737](https://github.com/tobymao/sqlglot/issues/8737) opened by [@DeepBuha06](https://github.com/DeepBuha06)*
+- [`4678f6c`](https://github.com/tobymao/sqlglot/commit/4678f6cb67662b0156d942f603eee816abbfec5a) - **optimizer**: preserve row references and aliases during TVF inference *(PR [#8723](https://github.com/tobymao/sqlglot/pull/8723) by [@georgesittas](https://github.com/georgesittas))*
+- [`8fb04fb`](https://github.com/tobymao/sqlglot/commit/8fb04fb3428d592ba785dfc75bff19cf29d6d198) - **sqlite**: normalize hexadecimal integer literals *(PR [#8744](https://github.com/tobymao/sqlglot/pull/8744) by [@georgesittas](https://github.com/georgesittas))*
+- [`b994fa7`](https://github.com/tobymao/sqlglot/commit/b994fa759bb85182ae4a589c1d01973829236047) - **optimizer**: don't fold decimal division by zero [CLAUDE] *(PR [#8748](https://github.com/tobymao/sqlglot/pull/8748) by [@Debasishhh](https://github.com/Debasishhh))*
+  - :arrow_lower_right: *fixes issue [#8747](https://github.com/tobymao/sqlglot/issues/8747) opened by [@Debasishhh](https://github.com/Debasishhh)*
+- [`dd71d36`](https://github.com/tobymao/sqlglot/commit/dd71d360f4ac49fc67b8a444b5c4c557e67298ea) - **duckdb**: keep the fraction of an interval written as one string [CLAUDE] *(PR [#8746](https://github.com/tobymao/sqlglot/pull/8746) by [@vidigoat](https://github.com/vidigoat))*
+- [`7600123`](https://github.com/tobymao/sqlglot/commit/76001235539919787dfd05923c90d39b97122773) - **optimizer**: resolve Snowflake token-splitter columns and BigQuery pivot rows *(PR [#8745](https://github.com/tobymao/sqlglot/pull/8745) by [@georgesittas](https://github.com/georgesittas))*
+- [`eb03a97`](https://github.com/tobymao/sqlglot/commit/eb03a9726f27ce33b2a899e2b70c52ba8e6de8ad) - **postgres**: transpile a case-insensitive regex flag to ~* [CLAUDE] *(PR [#8755](https://github.com/tobymao/sqlglot/pull/8755) by [@vidigoat](https://github.com/vidigoat))*
+- [`a9f1309`](https://github.com/tobymao/sqlglot/commit/a9f1309cc7116a79008c9a2b6ab0a06ce6b6f4c7) - **optimizer**: rename right operand references when aligning UNION BY NAME *(PR [#8749](https://github.com/tobymao/sqlglot/pull/8749) by [@treysp](https://github.com/treysp))*
+- [`e0ac5e4`](https://github.com/tobymao/sqlglot/commit/e0ac5e4e667146df9fce47fd3a449bdd19f4f8ea) - keep the last value of an exclusive series with a negative step [CLAUDE] *(PR [#8757](https://github.com/tobymao/sqlglot/pull/8757) by [@vidigoat](https://github.com/vidigoat))*
+- [`f872c44`](https://github.com/tobymao/sqlglot/commit/f872c442af585d773247f623cd5878d17cda6850) - **optimizer**: preserve filtered and grouped CROSS JOIN inputs [CODEX] *(PR [#8756](https://github.com/tobymao/sqlglot/pull/8756) by [@Seraph-x](https://github.com/Seraph-x))*
+  - :arrow_lower_right: *fixes issue [#8754](https://github.com/tobymao/sqlglot/issues/8754) opened by [@hutiefang76](https://github.com/hutiefang76)*
+- [`4aa189a`](https://github.com/tobymao/sqlglot/commit/4aa189a472639e7af442c8d794514ec3ab116995) - **duckdb**: decode the escapes of a Unicode string into its characters [CLAUDE] *(PR [#8758](https://github.com/tobymao/sqlglot/pull/8758) by [@vidigoat](https://github.com/vidigoat))*
+- [`184557f`](https://github.com/tobymao/sqlglot/commit/184557f70dd6b606e76fb2592bd096e4667450eb) - **oracle**: parse JSON_VALUE RETURNING clause *(PR [#8762](https://github.com/tobymao/sqlglot/pull/8762) by [@madarwish007](https://github.com/madarwish007))*
+  - :arrow_lower_right: *fixes issue [#8761](https://github.com/tobymao/sqlglot/issues/8761) opened by [@madarwish007](https://github.com/madarwish007)*
+- [`37f0963`](https://github.com/tobymao/sqlglot/commit/37f09630d14be44def882c8a2ef0666725a4d83a) - **optimizer**: GROUP BY ordinal interpretation in unnested subqueries *(PR [#8730](https://github.com/tobymao/sqlglot/pull/8730) by [@fivetran-kwoodbeck](https://github.com/fivetran-kwoodbeck))*
+- [`4d7421e`](https://github.com/tobymao/sqlglot/commit/4d7421ee43612043f11574d98227319d5e1f9855) - **optimizer**: keep CROSS JOIN with LIMIT 1 and split the LEFT JOIN row check *(PR [#8760](https://github.com/tobymao/sqlglot/pull/8760) by [@geooo109](https://github.com/geooo109))*
+  - :arrow_lower_right: *fixes issue [#8750](https://github.com/tobymao/sqlglot/issues/8750) opened by [@wjjc1017](https://github.com/wjjc1017)*
+- [`78b24a7`](https://github.com/tobymao/sqlglot/commit/78b24a744adb8d5dc0eb285f067087338b60f266) - **postgres**: transpile hex SHA2 hashes with ENCODE(..., 'hex') and parse SHA256 as a digest [CLAUDE] *(PR [#8764](https://github.com/tobymao/sqlglot/pull/8764) by [@vidigoat](https://github.com/vidigoat))*
+- [`c068aef`](https://github.com/tobymao/sqlglot/commit/c068aefd6997c5d272af86fb560ee25d8d5a10d6) - **parser**: copy DataType node for each cast in _replace_lambda *(PR [#8780](https://github.com/tobymao/sqlglot/pull/8780) by [@Pyhroff](https://github.com/Pyhroff))*
+  - :arrow_lower_right: *fixes issue [#8777](https://github.com/tobymao/sqlglot/issues/8777) opened by [@olmaks](https://github.com/olmaks)*
+- [`a8a7610`](https://github.com/tobymao/sqlglot/commit/a8a7610d5d4139fe14d2ea624652fdbdb5ba74b3) - **bigquery**: generate EXTRACT(DAYOFWEEK FROM ...) for DayOfWeek *(PR [#8778](https://github.com/tobymao/sqlglot/pull/8778) by [@JoeyTan21](https://github.com/JoeyTan21))*
+  - :arrow_lower_right: *fixes issue [#8767](https://github.com/tobymao/sqlglot/issues/8767) opened by [@JoeyTan21](https://github.com/JoeyTan21)*
+- [`d375836`](https://github.com/tobymao/sqlglot/commit/d3758367d4205bc728738110363e82edb3be211b) - **generator**: escape text before quoting it in generated SQL *(PR [#8779](https://github.com/tobymao/sqlglot/pull/8779) by [@tobymao](https://github.com/tobymao))*
+- [`8c5600a`](https://github.com/tobymao/sqlglot/commit/8c5600ad1fa322cbc11dcd93ee78cf5608efeda5) - **clickhouse**: parenthesize Contains under unary minus [CLAUDE] *(PR [#8782](https://github.com/tobymao/sqlglot/pull/8782) by [@ekut](https://github.com/ekut))*
+- [`908809d`](https://github.com/tobymao/sqlglot/commit/908809d3c2b756d0c4e259c55e8e61ca9579ebd8) - **sqlite**: parse REAL and FLOAT as 8-byte DOUBLE [CLAUDE] *(PR [#8784](https://github.com/tobymao/sqlglot/pull/8784) by [@kennywu411-dev](https://github.com/kennywu411-dev))*
+- [`bc1cd99`](https://github.com/tobymao/sqlglot/commit/bc1cd99a78974337a592662023e24231c2bf2a91) - **parser**: infinite loop issue *(PR [#8788](https://github.com/tobymao/sqlglot/pull/8788) by [@georgesittas](https://github.com/georgesittas))*
+  - :arrow_lower_right: *fixes issue [#8786](https://github.com/tobymao/sqlglot/issues/8786) opened by [@RaghavChamadiya](https://github.com/RaghavChamadiya)*
+- [`0e7b2da`](https://github.com/tobymao/sqlglot/commit/0e7b2da5d199388e84908aef13c53e535f1cf715) - **optimizer**: preserve grouping when removing redundant casts *(PR [#8787](https://github.com/tobymao/sqlglot/pull/8787) by [@georgesittas](https://github.com/georgesittas))*
+- [`f0b4347`](https://github.com/tobymao/sqlglot/commit/f0b4347b84b71bfd406e2fe831656a6d03351110) - **optimizer**: coerce `VALUES` column types across all rows *(PR [#8775](https://github.com/tobymao/sqlglot/pull/8775) by [@drzippie](https://github.com/drzippie))*
+- [`5859d71`](https://github.com/tobymao/sqlglot/commit/5859d7146aca266f3fdfbff6f11a8eeffcb4ee78) - **duckdb**: keep every path of a multi-path JSON_EXTRACT [CLAUDE] *(PR [#8771](https://github.com/tobymao/sqlglot/pull/8771) by [@vidigoat](https://github.com/vidigoat))*
+- [`bc7a5c4`](https://github.com/tobymao/sqlglot/commit/bc7a5c466e8dffe653c3fe04cb080cef1de1ad41) - **optimizer**: avoid unsafe `LEFT JOIN` elimination with `DISTINCT ON` *(PR [#8769](https://github.com/tobymao/sqlglot/pull/8769) by [@treysp](https://github.com/treysp))*
+  - :arrow_lower_right: *fixes issue [#8751](https://github.com/tobymao/sqlglot/issues/8751) opened by [@wjjc1017](https://github.com/wjjc1017)*
+- [`33c7209`](https://github.com/tobymao/sqlglot/commit/33c72094fbe6d05382e560e1026780372640f79b) - **optimizer**: exclude semi/anti join RHS sources from column resolution *(PR [#8763](https://github.com/tobymao/sqlglot/pull/8763) by [@fivetran-amrutabhimsenayachit](https://github.com/fivetran-amrutabhimsenayachit))*
+- [`634edd9`](https://github.com/tobymao/sqlglot/commit/634edd93bedaa5cb624098bb66fb99e07706f4ce) - **sqlite**: generate functions for temporal casts *(PR [#8793](https://github.com/tobymao/sqlglot/pull/8793) by [@georgesittas](https://github.com/georgesittas))*
+  - :arrow_lower_right: *fixes issue [#8791](https://github.com/tobymao/sqlglot/issues/8791) opened by [@kennywu411-dev](https://github.com/kennywu411-dev)*
+- [`ebd5fb4`](https://github.com/tobymao/sqlglot/commit/ebd5fb4d74f6e849ef989c2a81ab8e69e0395569) - **duckdb**: drop the SQL/JSON clauses of JSON_ARRAY and JSON_OBJECT [CLAUDE] *(PR [#8794](https://github.com/tobymao/sqlglot/pull/8794) by [@vidigoat](https://github.com/vidigoat))*
+- [`a357b7d`](https://github.com/tobymao/sqlglot/commit/a357b7d9be3d7c03ece213d53421d8a192b90a03) - **postgres**: support tablespace clauses in CREATE statements [CODEX] *(PR [#8792](https://github.com/tobymao/sqlglot/pull/8792) by [@georgesittas](https://github.com/georgesittas))*
+  - :arrow_lower_right: *fixes issue [#8776](https://github.com/tobymao/sqlglot/issues/8776) opened by [@cmodiano](https://github.com/cmodiano)*
+- [`31918e9`](https://github.com/tobymao/sqlglot/commit/31918e9a2f249bb5cda263e644eb9c1b96be0c6d) - **duckdb**: preserve quoted variable names *(PR [#8795](https://github.com/tobymao/sqlglot/pull/8795) by [@georgesittas](https://github.com/georgesittas))*
+- [`c5fe540`](https://github.com/tobymao/sqlglot/commit/c5fe5404f379d84798610ab23fb94980f26d2ff2) - **snowflake**: support MERGE ALL BY NAME *(PR [#8797](https://github.com/tobymao/sqlglot/pull/8797) by [@georgesittas](https://github.com/georgesittas))*
+  - :arrow_lower_right: *fixes issue [#8783](https://github.com/tobymao/sqlglot/issues/8783) opened by [@sai-sakalam](https://github.com/sai-sakalam)*
+- [`114999a`](https://github.com/tobymao/sqlglot/commit/114999ae6787cbdb7ea630710b82ad2849711e08) - **optimizer**: avoid nested aggregates when unnesting scalar subqueries [CODEX] *(PR [#8798](https://github.com/tobymao/sqlglot/pull/8798) by [@georgesittas](https://github.com/georgesittas))*
+  - :arrow_lower_right: *fixes issue [#8790](https://github.com/tobymao/sqlglot/issues/8790) opened by [@kennywu411-dev](https://github.com/kennywu411-dev)*
+- [`fb36fb7`](https://github.com/tobymao/sqlglot/commit/fb36fb7112ad71356fe81fd2cce9aa4733679e1d) - **optimizer**: annotate struct-valued subqueries and row references *(PR [#8694](https://github.com/tobymao/sqlglot/pull/8694) by [@georgesittas](https://github.com/georgesittas))*
+
+### :zap: Performance Improvements
+- [`a2a8bb4`](https://github.com/tobymao/sqlglot/commit/a2a8bb4209e188c0ae78d9809ec3e27666c00f16) - **athena**: avoid tokenizing DML twice [CLAUDE] *(PR [#8709](https://github.com/tobymao/sqlglot/pull/8709) by [@masiljangajji](https://github.com/masiljangajji))*
+  - :arrow_lower_right: *addresses issue [#8708](https://github.com/tobymao/sqlglot/issues/8708) opened by [@masiljangajji](https://github.com/masiljangajji)*
+- [`7c4e124`](https://github.com/tobymao/sqlglot/commit/7c4e124f21225a4b60d790cb6938d64bf2778d4a) - **optimizer**: cache positional column alias mappings [CODEX] *(PR [#8735](https://github.com/tobymao/sqlglot/pull/8735) by [@georgesittas](https://github.com/georgesittas))*
+
+### :recycle: Refactors
+- [`65eb0f2`](https://github.com/tobymao/sqlglot/commit/65eb0f229a5d7af5e4d3ff9084d79362e5fc899a) - **duckdb**: drop unused Lambda colon arg *(PR [#8721](https://github.com/tobymao/sqlglot/pull/8721) by [@VaggelisD](https://github.com/VaggelisD))*
+- [`2bf94d9`](https://github.com/tobymao/sqlglot/commit/2bf94d96359db584862dcf20b358f34cabc1661d) - **optimizer**: expand struct column paths using inferred types *(PR [#8695](https://github.com/tobymao/sqlglot/pull/8695) by [@georgesittas](https://github.com/georgesittas))*
+- [`9d2966e`](https://github.com/tobymao/sqlglot/commit/9d2966ed3f8a7f0d370846dbae1579b40ddc2924) - **optimizer**: defer scope annotation until qualification is complete *(PR [#8696](https://github.com/tobymao/sqlglot/pull/8696) by [@georgesittas](https://github.com/georgesittas))*
+
+### :wrench: Chores
+- [`9fedfa5`](https://github.com/tobymao/sqlglot/commit/9fedfa5f5ecb94f4ca290ee6373b1dd4feb120e2) - fix comments for insert table *(commit by [@geooo109](https://github.com/geooo109))*
+- [`3377b95`](https://github.com/tobymao/sqlglot/commit/3377b95795837881a1ffe86c0ea6965084b195d7) - minimize insert table parser *(commit by [@geooo109](https://github.com/geooo109))*
+- [`9075977`](https://github.com/tobymao/sqlglot/commit/90759771ddf61d8b53d729712c53870a3317f5ce) - suppress unsupported property warning in tests *(PR [#8724](https://github.com/tobymao/sqlglot/pull/8724) by [@treysp](https://github.com/treysp))*
+
+
 ## [v30.24.0] - 2026-10-20
 ### :boom: BREAKING CHANGES
 - due to [`094509a`](https://github.com/tobymao/sqlglot/commit/094509aad83fff575f1b4d198e4d03c21c432e58) - merge SortArray into ArraySort with explicit NULL ordering *(PR [#8643](https://github.com/tobymao/sqlglot/pull/8643) by [@tobymao](https://github.com/tobymao))*:
@@ -16385,3 +16492,4 @@ pip install "sqlglot[c]"   # compiled — faster, but no subclassing
 [v30.22.0]: https://github.com/tobymao/sqlglot/compare/v30.21.0...v30.22.0
 [v30.23.0]: https://github.com/tobymao/sqlglot/compare/v30.22.0...v30.23.0
 [v30.24.0]: https://github.com/tobymao/sqlglot/compare/v30.23.0...v30.24.0
+[v30.25.0]: https://github.com/tobymao/sqlglot/compare/v30.24.0...v30.25.0

{sqlglot-30.25.0 → sqlglot-30.26.0}/Makefile

@@ -28,8 +28,8 @@ install-dev:
 	@if [ -f sqlglot-integration-tests/Makefile ]; then \
 		$(MAKE) -C sqlglot-integration-tests install; \
 	fi
+	$(PIP) uninstall $(if $(UV),,-y) sqlglotc 2>/dev/null || true
 	$(PIP) install -e ".[dev]"
-	$(PIP) uninstall -y sqlglotc 2>/dev/null || true
 	@if ! command -v gh >/dev/null 2>&1; then \
 		echo ""; \
 		echo "gh (GitHub CLI) is not installed. It is needed to auto-create PRs for integration tests."; \

{sqlglot-30.25.0/sqlglot.egg-info → sqlglot-30.26.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sqlglot
-Version: 30.25.0
+Version: 30.26.0
 Summary: An easily customizable SQL parser and transpiler
 Author-email: Toby Mao <toby.mao@gmail.com>
 License-Expression: MIT
@@ -21,7 +21,7 @@ Provides-Extra: dev
 Requires-Dist: duckdb>=0.6; extra == "dev"
 Requires-Dist: mypy>=2.4.0; python_version >= "3.10" and extra == "dev"
 Requires-Dist: mypy; python_version < "3.10" and extra == "dev"
-Requires-Dist: setuptools_scm; extra == "dev"
+Requires-Dist: setuptools_scm>=10.2; extra == "dev"
 Requires-Dist: pandas; extra == "dev"
 Requires-Dist: pandas-stubs; extra == "dev"
 Requires-Dist: python-dateutil; extra == "dev"
@@ -34,10 +34,10 @@ Requires-Dist: types-pytz; extra == "dev"
 Requires-Dist: typing_extensions; extra == "dev"
 Requires-Dist: pyperf; extra == "dev"
 Provides-Extra: c
-Requires-Dist: sqlglotc==30.25.0; python_version >= "3.10" and extra == "c"
+Requires-Dist: sqlglotc==30.26.0; python_version >= "3.10" and extra == "c"
 Provides-Extra: rs
 Requires-Dist: sqlglotrs==0.13.0; extra == "rs"
-Requires-Dist: sqlglotc==30.25.0; python_version >= "3.10" and extra == "rs"
+Requires-Dist: sqlglotc==30.26.0; python_version >= "3.10" and extra == "rs"
 Dynamic: license-file
 Dynamic: provides-extra
 

{sqlglot-30.25.0 → sqlglot-30.26.0}/pyproject.toml

@@ -23,7 +23,7 @@ Repository = "https://github.com/tobymao/sqlglot"
 Issues = "https://github.com/tobymao/sqlglot/issues"
 
 [build-system]
-requires = ["setuptools >= 61.0", "setuptools_scm"]
+requires = ["setuptools >= 61.0", "setuptools_scm >= 10.2"]
 build-backend = "setuptools.build_meta"
 
 [tool.setuptools]

{sqlglot-30.25.0 → sqlglot-30.26.0}/setup.py

@@ -11,7 +11,7 @@ setup(
             # for py3.10+; on py3.9 just use upstream mypy for type checking.
             "mypy >= 2.4.0; python_version >= '3.10'",
             "mypy; python_version < '3.10'",
-            "setuptools_scm",
+            "setuptools_scm >= 10.2",
             "pandas",
             "pandas-stubs",
             "python-dateutil",

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '30.25.0'
-__version_tuple__ = version_tuple = (30, 25, 0)
+__version__ = version = '30.26.0'
+__version_tuple__ = version_tuple = (30, 26, 0)
 
-__commit_id__ = commit_id = 'g9d2966ed3'
+__commit_id__ = commit_id = 'gb0842a1ee'

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/dialects/bigquery.py

@@ -211,6 +211,7 @@ class BigQuery(Dialect):
             "TIMESTAMP": TokenType.TIMESTAMPTZ,
             "WHILE": TokenType.COMMAND,
         }
+        KEYWORDS.pop("ISNULL")
         KEYWORDS.pop("DIV")
         KEYWORDS.pop("VALUES")
         KEYWORDS.pop("/*+")

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/dialects/dialect.py

@@ -42,7 +42,15 @@ DATE_ADD_OR_DIFF = t.Union[
     exp.TsOrDsAdd,
     exp.TsOrDsDiff,
 ]
-DATE_ADD_OR_SUB = t.Union[exp.DateAdd, exp.TsOrDsAdd, exp.DateSub]
+DATE_ADD_OR_SUB = t.Union[
+    exp.DateAdd,
+    exp.DateSub,
+    exp.DatetimeAdd,
+    exp.DatetimeSub,
+    exp.TimestampAdd,
+    exp.TimestampSub,
+    exp.TsOrDsAdd,
+]
 JSON_EXTRACT_TYPE = t.Union[
     exp.JSONExtract, exp.JSONExtractScalar, exp.JSONBExtract, exp.JSONBExtractScalar
 ]
@@ -804,6 +812,12 @@ class Dialect(metaclass=_Dialect):
     must be explicitly specified.
     """
 
+    INTERSECT_BINDS_TIGHTER_THAN_UNION_AND_EXCEPT = True
+    """
+    Whether INTERSECT binds more tightly than UNION and EXCEPT when a set operation chain isn't
+    parenthesized. If False, all set operations have equal precedence and associate left to right.
+    """
+
     CREATABLE_KIND_MAPPING: dict[str, str] = {}
     """
     Helper for dialects that use a different name for the same creatable kind. For example, the Clickhouse
@@ -2220,13 +2234,23 @@ def date_delta_sql(name: str, cast: bool = False) -> t.Callable[[Generator, DATE
     return _delta_sql
 
 
+def date_delta_to_binary_sql(self: Generator, expression: exp.Expr, binary: exp.Binary) -> str:
+    parent = expression.parent
+    if isinstance(parent, (exp.Binary, exp.Unary, exp.AtTimeZone)) and not (
+        isinstance(parent, exp.Paren)
+        or isinstance(parent, (exp.Add, exp.Sub))
+        and parent.this is expression
+    ):
+        return self.sql(exp.paren(binary, copy=False))
+    return self.sql(binary)
+
+
 def date_delta_to_binary_interval_op(
     cast: bool = True,
 ) -> t.Callable[[Generator, DATETIME_DELTA], str]:
     def date_delta_to_binary_interval_op_sql(self: Generator, expression: DATETIME_DELTA) -> str:
         this = expression.this
         unit = unit_to_var(expression)
-        op = "+" if isinstance(expression, DATETIME_ADD) else "-"
 
         to_type: exp.DATA_TYPE | None = None
         if cast:
@@ -2236,7 +2260,15 @@ def date_delta_to_binary_interval_op(
                 # Cast string literals (i.e function parameters) to the appropriate type for +/- interval to work
                 to_type = (
                     exp.DType.DATETIME
-                    if isinstance(expression, (exp.DatetimeAdd, exp.DatetimeSub))
+                    if isinstance(
+                        expression,
+                        (
+                            exp.DatetimeAdd,
+                            exp.DatetimeSub,
+                            exp.TimestampAdd,
+                            exp.TimestampSub,
+                        ),
+                    )
                     else exp.DType.DATE
                 )
 
@@ -2245,7 +2277,11 @@ def date_delta_to_binary_interval_op(
         expr = expression.expression
         interval = expr if isinstance(expr, exp.Interval) else exp.Interval(this=expr, unit=unit)
 
-        return f"{self.sql(this)} {op} {self.sql(interval)}"
+        return date_delta_to_binary_sql(
+            self,
+            expression,
+            this + interval if isinstance(expression, DATETIME_ADD) else this - interval,
+        )
 
     return date_delta_to_binary_interval_op_sql
 

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/dialects/hive.py

@@ -24,6 +24,7 @@ class Hive(Dialect):
     ARRAY_AGG_INCLUDES_NULLS = None
     REGEXP_EXTRACT_DEFAULT_GROUP = 1
     ALTER_TABLE_SUPPORTS_CASCADE = True
+    INTERSECT_BINDS_TIGHTER_THAN_UNION_AND_EXCEPT = False
 
     # https://spark.apache.org/docs/latest/sql-ref-identifier.html#description
     NORMALIZATION_STRATEGY = NormalizationStrategy.CASE_INSENSITIVE

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/dialects/oracle.py

@@ -21,6 +21,7 @@ class Oracle(Dialect):
     ON_CONDITION_EMPTY_BEFORE_ERROR = False
     ALTER_TABLE_ADD_REQUIRED_FOR_EACH_COLUMN = False
     DISABLES_ALIAS_REF_EXPANSION = True
+    INTERSECT_BINDS_TIGHTER_THAN_UNION_AND_EXCEPT = False
 
     # See section 8: https://docs.oracle.com/cd/A97630_01/server.920/a96540/sql_elements9a.htm
     NORMALIZATION_STRATEGY = NormalizationStrategy.UPPERCASE

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/dialects/snowflake.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+from itertools import product
+
 from sqlglot import exp, jsonpath, tokens
 from sqlglot.dialects.dialect import (
     Dialect,
@@ -47,68 +49,47 @@ class Snowflake(Dialect):
     INITCAP_DEFAULT_DELIMITER_CHARS = ' \t\n\r\f\v!?@"^#$&~_,.:;+\\-*%/|\\[\\](){}<>'
 
     INVERSE_TIME_MAPPING = {
+        "%d": "DD",
+        "%-d": "DD",
+        "%a": "DY",
         "T": "T",  # in TIME_MAPPING we map '"T"' with the double quotes to 'T', and we want to prevent 'T' from being mapped back to '"T"' so that 'AUTO' doesn't become 'AU"T"O'
     }
 
     TIME_MAPPING = {
-        "YYYY": "%Y",
-        "yyyy": "%Y",
-        "YY": "%y",
-        "yy": "%y",
-        "MMMM": "%B",
-        "mmmm": "%B",
-        "MON": "%b",
-        "mon": "%b",
-        "MM": "%m",
-        "mm": "%m",
-        "DD": "%d",
-        "dd": "%-d",
-        "DY": "%a",
-        "dy": "%w",
-        "HH24": "%H",
-        "hh24": "%H",
-        "HH12": "%I",
-        "hh12": "%I",
-        "MI": "%M",
-        "mi": "%M",
-        "SS": "%S",
-        "ss": "%S",
-        "FF": "%f_nine",  # %f_ internal representation with precision specified
-        "ff": "%f_nine",
-        "FF0": "%f_zero",
-        "ff0": "%f_zero",
-        "FF1": "%f_one",
-        "ff1": "%f_one",
-        "FF2": "%f_two",
-        "ff2": "%f_two",
-        "FF3": "%f_three",
-        "ff3": "%f_three",
-        "FF4": "%f_four",
-        "ff4": "%f_four",
-        "FF5": "%f_five",
-        "ff5": "%f_five",
-        "FF6": "%f",
-        "ff6": "%f",
-        "FF7": "%f_seven",
-        "ff7": "%f_seven",
-        "FF8": "%f_eight",
-        "ff8": "%f_eight",
-        "FF9": "%f_nine",
-        "ff9": "%f_nine",
-        "TZHTZM": "%z",
-        "tzhtzm": "%z",
-        "TZH:TZM": "%:z",  # internal representation for ±HH:MM
-        "tzh:tzm": "%:z",
-        "TZH": "%-z",  # internal representation ±HH
-        "tzh": "%-z",
-        '"T"': "T",  # remove the optional double quotes around the separator between the date and time
-        # Seems like Snowflake treats AM/PM in the format string as equivalent,
-        # only the time (stamp) value's AM/PM affects the output
-        "AM": "%p",
-        "am": "%p",
-        "PM": "%p",
-        "pm": "%p",
-    }
+        "".join(chars): value
+        for key, value in {
+            "YYYY": "%Y",
+            "YY": "%y",
+            "MMMM": "%B",
+            "MON": "%b",
+            "MM": "%m",
+            "DD": "%d",
+            "DY": "%a",
+            "HH24": "%H",
+            "HH12": "%I",
+            "MI": "%M",
+            "SS": "%S",
+            "FF": "%f_nine",  # %f_ internal representation with precision specified
+            "FF0": "%f_zero",
+            "FF1": "%f_one",
+            "FF2": "%f_two",
+            "FF3": "%f_three",
+            "FF4": "%f_four",
+            "FF5": "%f_five",
+            "FF6": "%f",
+            "FF7": "%f_seven",
+            "FF8": "%f_eight",
+            "FF9": "%f_nine",
+            "TZHTZM": "%z",
+            "TZH:TZM": "%:z",  # internal representation for ±HH:MM
+            "TZH": "%-z",  # internal representation ±HH
+            # Seems like Snowflake treats AM/PM in the format string as equivalent,
+            # only the time (stamp) value's AM/PM affects the output
+            "AM": "%p",
+            "PM": "%p",
+        }.items()
+        for chars in product(*zip(key.upper(), key.lower()))
+    } | {'"T"': "T"}
 
     DATE_PART_MAPPING = {
         **Dialect.DATE_PART_MAPPING,
@@ -189,6 +170,8 @@ class Snowflake(Dialect):
             "FLOAT": TokenType.DOUBLE,
         }
         KEYWORDS.pop("/*+")
+        KEYWORDS.pop("GLOB")
+        KEYWORDS.pop("RETURNING")
 
         SINGLE_TOKENS = {
             **tokens.Tokenizer.SINGLE_TOKENS,

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/dialects/spark2.py

@@ -9,6 +9,7 @@ from sqlglot.typing.spark2 import EXPRESSION_METADATA
 
 class Spark2(Hive):
     ALTER_TABLE_SUPPORTS_CASCADE = False
+    INTERSECT_BINDS_TIGHTER_THAN_UNION_AND_EXCEPT = True
 
     EXPRESSION_METADATA = EXPRESSION_METADATA.copy()
 

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/dialects/sqlite.py

@@ -21,6 +21,7 @@ class SQLite(Dialect):
     USING_COLUMN_ORDER = UsingColumnOrder.IN_PLACE
     CONCAT_COALESCE = True
     CONCAT_WS_COALESCE = True
+    INTERSECT_BINDS_TIGHTER_THAN_UNION_AND_EXCEPT = False
 
     class Tokenizer(tokens.Tokenizer):
         IDENTIFIERS = ['"', ("[", "]"), "`"]

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/expressions/core.py

@@ -2108,6 +2108,10 @@ class Filter(Expression):
     arg_types = {"this": True, "expression": True}
 
 
+class Per(Expression):
+    arg_types = {"this": True, "expressions": True}
+
+
 class Check(Expression):
     pass
 

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/expressions/properties.py

@@ -107,6 +107,10 @@ class CopyGrantsProperty(Property):
     arg_types = {}
 
 
+class CopyTagsProperty(Property):
+    arg_types = {}
+
+
 class DataBlocksizeProperty(Property):
     arg_types = {
         "size": False,

{sqlglot-30.25.0 → sqlglot-30.26.0}/sqlglot/generator.py

@@ -169,6 +169,7 @@ class Generator:
             "CONVERT", e.this, e.args["dest"], e.args.get("source")
         ),
         exp.CopyGrantsProperty: lambda *_: "COPY GRANTS",
+        exp.CopyTagsProperty: lambda *_: "COPY TAGS",
         exp.CredentialsProperty: lambda self, e: (
             f"CREDENTIALS=({self.expressions(e, 'expressions', sep=' ')})"
         ),
@@ -731,6 +732,7 @@ class Generator:
         exp.CollateProperty: exp.Properties.Location.POST_SCHEMA,
         exp.ComputeProperty: exp.Properties.Location.POST_CREATE,
         exp.CopyGrantsProperty: exp.Properties.Location.POST_SCHEMA,
+        exp.CopyTagsProperty: exp.Properties.Location.POST_SCHEMA,
         exp.Cluster: exp.Properties.Location.POST_SCHEMA,
         exp.ClusteredByProperty: exp.Properties.Location.POST_SCHEMA,
         exp.ClusterProperty: exp.Properties.Location.POST_SCHEMA,
@@ -1964,19 +1966,63 @@ class Generator:
                 return self.sql(select)
 
         sqls: list[str] = []
-        stack: list[str | exp.Expr] = [expression]
+        stack: list[str | exp.Expr | int | None] = [expression]
+        depth = 0
+        open_parens = ""
 
         while stack:
             node = stack.pop()
 
-            if isinstance(node, exp.SetOperation):
+            if isinstance(node, int):
+                # Closes a plain grouping opened below, restoring its depth
+                depth = node
+                sqls[-1] += self.sep("") + self.indent(")", level=depth, pad=0)
+            elif isinstance(node, exp.SetOperation):
                 stack.append(node.expression)
                 stack.append(
                     self.maybe_comment(
                         self.set_operation(node), comments=node.comments, separated=True
                     )
                 )
-                stack.append(node.this)
+                left = node.this
+                if (
+                    not self.SET_OP_PARENTHESIZED_OPERANDS
+                    and not self.dialect.INTERSECT_BINDS_TIGHTER_THAN_UNION_AND_EXCEPT
+                ):
+                    # Left grouping is implicit, so skip the `SELECT * FROM (...)` wrapper:
+                    # it renames duplicate columns and deepens recursion. Inner CTEs and
+                    # modifiers still need the wrapper.
+                    wrappers: list[exp.Subquery] = []
+                    inner = left
+                    while isinstance(inner, exp.Subquery) and inner.is_wrapper:
+                        wrappers.append(inner)
+                        inner = inner.this
+
+                    if (
+                        isinstance(inner, exp.SetOperation)
+                        and not inner.args.get("with_")
+                        and not any(inner.args.get(key) for key in exp.QUERY_MODIFIERS)
+                    ):
+                        for wrapper in wrappers:
+                            inner.add_comments(wrapper.pop_comments())
+                        left = inner
+                elif (
+                    self.SET_OP_PARENTHESIZED_OPERANDS
+                    and isinstance(left, exp.Subquery)
+                    and left.is_wrapper
+                    and not left.comments
+                    and isinstance(left.this, exp.SetOperation)
+                    and not left.this.args.get("with_")
+                    and not any(left.this.args.get(key) for key in exp.QUERY_MODIFIERS)
+                ):
+                    # Write a plain grouping's parentheses here rather than recursing through
+                    # subquery_sql: left-to-right chains nest one grouping per INTERSECT, and
+                    # long ones would exhaust the stack. The depth pushed here marks the close.
+                    open_parens += self.indent("(", level=depth, pad=0) + self.sep("")
+                    stack.append(depth)
+                    depth += 1
+                    left = left.this
+                stack.append(left)
             else:
                 if (
                     not self.SET_OP_LIMITS
@@ -1986,7 +2032,14 @@ class Generator:
                     node = node.subquery(copy=False)
                     if not self.SET_OP_PARENTHESIZED_OPERANDS:
                         node = exp.select("*").from_(node, copy=False)
-                sqls.append(self.sql(node))
+                elif (
+                    not self.SET_OP_PARENTHESIZED_OPERANDS
+                    and isinstance(node, exp.Subquery)
+                    and isinstance(node.unnest(), exp.SetOperation)
+                ):
+                    node = exp.select("*").from_(node, copy=False)
+                sqls.append(open_parens + self.indent(self.sql(node), level=depth, pad=0))
+                open_parens = ""
 
         this = self.sep().join(sqls)
         this = self.query_modifiers(expression, this)
@@ -2014,6 +2067,9 @@ class Generator:
         where = self.sql(expression, "expression").strip()
         return f"{this} FILTER({where})"
 
+    def per_sql(self, expression: exp.Per) -> str:
+        return f"{self.sql(expression, 'this')} PER ({self.expressions(expression, flat=True)})"
+
     def hint_sql(self, expression: exp.Hint) -> str:
         if not self.QUERY_HINTS:
             self.unsupported("Hints are not supported")