From f50f6b8b747005d75db5832954571da947864323 Mon Sep 17 00:00:00 2001
From: Matt Kornfield <mkornfield@nvidia.com>
Date: Thu, 21 May 2026 09:38:07 -0700
Subject: [PATCH 1/2] chore: bump sqlfluff to 4.1.0

---
 packages/data-designer-engine/pyproject.toml | 2 +-
 uv.lock                                      | 9 ++++-----
 2 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/packages/data-designer-engine/pyproject.toml b/packages/data-designer-engine/pyproject.toml
index 87f19785b..c9fe9e225 100644
--- a/packages/data-designer-engine/pyproject.toml
+++ b/packages/data-designer-engine/pyproject.toml
@@ -53,7 +53,7 @@ dependencies = [
     "python-multipart>=0.0.27,<1", # 0.0.27 fixes security advisory pulled in by mcp
     "ruff>=0.14.10,<1",
     "scipy>=1.11.0,<2",
-    "sqlfluff>=3.2.0,<4",
+    "sqlfluff>=4.1.0,<5",
     "tiktoken>=0.8.0,<1",
 ]
 
diff --git a/uv.lock b/uv.lock
index 123ac8e8a..94e429aa8 100644
--- a/uv.lock
+++ b/uv.lock
@@ -909,7 +909,7 @@ requires-dist = [
     { name = "python-multipart", specifier = ">=0.0.27,<1" },
     { name = "ruff", specifier = ">=0.14.10,<1" },
     { name = "scipy", specifier = ">=1.11.0,<2" },
-    { name = "sqlfluff", specifier = ">=3.2.0,<4" },
+    { name = "sqlfluff", specifier = ">=4.1.0,<5" },
     { name = "tiktoken", specifier = ">=0.8.0,<1" },
 ]
 
@@ -4634,7 +4634,7 @@ wheels = [
 
 [[package]]
 name = "sqlfluff"
-version = "3.5.0"
+version = "4.2.1"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "chardet" },
@@ -4644,16 +4644,15 @@ dependencies = [
     { name = "jinja2" },
     { name = "pathspec" },
     { name = "platformdirs" },
-    { name = "pytest" },
     { name = "pyyaml" },
     { name = "regex" },
     { name = "tblib" },
     { name = "tomli", marker = "python_full_version < '3.11'" },
     { name = "tqdm" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/4c/a8/d3dc6c510cc3bba9abbf7a3052a96d5ce6771b71dda141846003fa37277a/sqlfluff-3.5.0.tar.gz", hash = "sha256:2d0a546078ffb021de7021b9a6c2a50e5eef590daa820d5f1b082d24a1d5e1d4", size = 921199, upload-time = "2025-10-18T19:33:07.778Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/b4/a1/3430aebc4fae35d7e466e793b5da2f36c2245af092311520dc1d3d3146d6/sqlfluff-4.2.1.tar.gz", hash = "sha256:32f43fbf6721e57f1a5a87d71df0d94b84ecba6ed65727266c7fa60991110fb9", size = 1013384, upload-time = "2026-05-14T21:15:37.161Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/47/d5/83c3eacdd6c3249fb5f8a0b5612ab10b661862e0df869951f45fd837448d/sqlfluff-3.5.0-py3-none-any.whl", hash = "sha256:6e5fb7a0c491676ded68912245fc0627e88f8b0e6290bd4b54a65ce735f69716", size = 921597, upload-time = "2025-10-18T19:33:05.839Z" },
+    { url = "https://files.pythonhosted.org/packages/5f/55/8830f3204939cc965c72680bfae99b0f3fd6c16bffdaf79372ad0a3d1ca6/sqlfluff-4.2.1-py3-none-any.whl", hash = "sha256:ea84f196c41f45df40a851b0881cb3fbb660570e07acbbfd304ff4e9b893424d", size = 1002493, upload-time = "2026-05-14T21:15:35.582Z" },
 ]
 
 [[package]]

From a3e01c11c6e526e54f2e37e8278eafb5942dc267 Mon Sep 17 00:00:00 2001
From: Matt Kornfield <mkornfield@nvidia.com>
Date: Fri, 22 May 2026 10:29:28 -0700
Subject: [PATCH 2/2] chore: add some sql format unit tests

Signed-off-by: Matt Kornfield <mkornfield@nvidia.com>
---
 .../tests/engine/validators/test_sql.py       | 20 +++++++++++++++----
 1 file changed, 16 insertions(+), 4 deletions(-)

diff --git a/packages/data-designer-engine/tests/engine/validators/test_sql.py b/packages/data-designer-engine/tests/engine/validators/test_sql.py
index 911ce1f4d..21b18dbe8 100644
--- a/packages/data-designer-engine/tests/engine/validators/test_sql.py
+++ b/packages/data-designer-engine/tests/engine/validators/test_sql.py
@@ -9,10 +9,22 @@
 from data_designer.config.validator_params import CodeValidatorParams
 from data_designer.engine.validators.sql import SQLValidator
 
-
-def test_valid_ansi_sql_code() -> None:
-    sql_validator = SQLValidator(CodeValidatorParams(code_lang=CodeLang.SQL_ANSI))
-    code = "SELECT category, COUNT(*) as total_incidents FROM security_incidents_2 GROUP BY category;"
+VALID_SQL_BY_DIALECT = [
+    (
+        CodeLang.SQL_ANSI,
+        "SELECT category, COUNT(*) AS total_incidents FROM security_incidents_2 GROUP BY category;",
+    ),
+    (CodeLang.SQL_SQLITE, "SELECT sqlite_version() AS version;"),
+    (CodeLang.SQL_TSQL, "SELECT TOP 1 name FROM sys.objects;"),
+    (CodeLang.SQL_BIGQUERY, "SELECT * EXCEPT(sensitive_column) FROM `project.dataset.table`;"),
+    (CodeLang.SQL_MYSQL, "SELECT `name` FROM users LIMIT 1;"),
+    (CodeLang.SQL_POSTGRES, "SELECT 1::int AS value;"),
+]
+
+
+@pytest.mark.parametrize(("code_lang", "code"), VALID_SQL_BY_DIALECT)
+def test_valid_sql_code_for_supported_dialects(code_lang: CodeLang, code: str) -> None:
+    sql_validator = SQLValidator(CodeValidatorParams(code_lang=code_lang))
     result = sql_validator.run_validation([{"sql": code}])
     assert result.data[0].is_valid
     assert result.data[0].error_messages == ""