feat(scenario): map scenario column types with SQLQuack's type parsers by kvlonge · Pull Request #1364 · sqlstack-io/sqlbuild

kvlonge · 2026-11-08T04:26:02Z

Why

Scenario capture still asked Polyglot (parse_data_type) to read warehouse column types. This replaces it with SQLQuack's type parsers, one of the last two runtime uses of the Polyglot package.

Changes

One native entry point, parse_warehouse_type(spelling, dialect). Postgres uses sqlquack-syntax-postgres; Snowflake and BigQuery use the cast target of CAST(NULL AS <type>). Other dialects return None.
columns.py: _local_type_with_polyglot and its four helpers are replaced by a walk of the SQLQuack tree. DuckDB and MotherDuck spellings pass through unchanged.
Databricks ARRAY/MAP/STRUCT map recursively to DuckDB T[], MAP(K, V) and STRUCT(a T). An unmappable element is refused with X605 instead of becoming JSON. SQL Server's date, time, uuid, bit and money types get their own pre-mapper. The fallback keeps integer and 4-byte float widths and contains no Polyglot.
The SQLQuack rev moves 4d0c6bc5 → 98eefae6, and sqlquack-syntax-bigquery is added. Fingerprints are unchanged because the tokenizer and scanner sources are identical. The next SQLQuack commit refuses FROM __ref(...) placeholders on Snowflake, so this rev stops before it.
import_polyglot.py is deleted. The local-SQL path and the dependency drop follow separately.

Verification

Before/after over 156 dialect and type pairs: the capture tests, SQL Server's information_schema names and float spellings. 31 rows change, each named in the latest table comment: spacing, TEXT→VARCHAR, BigQuery INTEGER→BIGINT (INT64), Snowflake VARIANT→JSON, DuckDB spellings passing through, and SQL Server types that were invalid in DuckDB.
Scenario unit and integration tests: 339 passed. Scenario E2E (16 files): 120 passed.
cargo clippy -D warnings, cargo fmt, ruff, ty, fensu, and the native stub, boundary and layering checks pass.

Local scenario capture no longer asks Polyglot to read warehouse column types. Postgres types go through sqlquack-syntax-postgres's type parser, and Snowflake and BigQuery types through their SQLQuack cast-target parse, behind one native function (parse_warehouse_type). DuckDB and MotherDuck spellings pass through unchanged. Databricks nested types map recursively to DuckDB's LIST, MAP and STRUCT, and a nested spelling that cannot be mapped is refused with a named message instead of falling back to JSON. Other types keep their integer width or fall back to VARCHAR.

kvlonge · 2026-11-08T04:32:57Z

Before/after type mapping for every dialect and type pair in the scenario capture tests (Polyglot on main vs SQLQuack here). 13 of 116 rows change.

Dialect	Warehouse type	Before (Polyglot)	After (SQLQuack)	Change
bigquery	`ARRAY<INT64>`	`BIGINT[]`	`BIGINT[]`
bigquery	`ARRAY<STRUCT<a INT64, b STRING>>`	`STRUCT(a BIGINT, b TEXT)[]`	`STRUCT(a BIGINT, b TEXT)[]`
bigquery	`BIGNUMERIC`	`DECIMAL(38, 5)`	`DECIMAL(38, 5)`
bigquery	`BIGNUMERIC(38,9)`	`DECIMAL(38, 9)`	`DECIMAL(38, 9)`
bigquery	`BOOL`	`BOOLEAN`	`BOOLEAN`
bigquery	`BYTES`	`VARCHAR`	`VARCHAR`
bigquery	`BYTES(10)`	`VARCHAR`	`VARCHAR`
bigquery	`DATE`	`DATE`	`DATE`
bigquery	`DATETIME`	`TIMESTAMP`	`TIMESTAMP`
bigquery	`FLOAT64`	`DOUBLE`	`DOUBLE`
bigquery	`GEOGRAPHY`	`VARCHAR`	`VARCHAR`
bigquery	`INT64`	`BIGINT`	`BIGINT`
bigquery	`INTEGER`	`INT`	`BIGINT`	BigQuery INTEGER is INT64
bigquery	`JSON`	`JSON`	`JSON`
bigquery	`NUMERIC`	`DECIMAL`	`DECIMAL`
bigquery	`NUMERIC(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10, 2)`
bigquery	`RANGE<DATE>`	`VARCHAR`	`VARCHAR`
bigquery	`RANGE<DATETIME>`	`VARCHAR`	`VARCHAR`
bigquery	`RANGE<TIMESTAMP>`	`VARCHAR`	`VARCHAR`
bigquery	`STRING`	`TEXT`	`TEXT`
bigquery	`STRUCT<a INT64>`	`STRUCT(a BIGINT)`	`STRUCT(a BIGINT)`
bigquery	`TIMESTAMP`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
databricks	`ARRAY<INT>`	`INT[]`	`INT[]`
databricks	`DECIMAL(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10,2)`	spacing only
databricks	`GEOGRAPHY`	`VARCHAR`	`VARCHAR`
databricks	`GEOMETRY`	`VARCHAR`	`VARCHAR`
databricks	`INTERVAL YEAR`	`INTERVAL YEAR`	`INTERVAL YEAR`
databricks	`MAP<STRING, ARRAY<INT>>`	`MAP(TEXT, INT[])`	`MAP(VARCHAR, INT[])`	TEXT and VARCHAR are the same DuckDB type
databricks	`MAP<STRING, INT>`	`MAP(TEXT, INT)`	`MAP(VARCHAR, INT)`	TEXT and VARCHAR are the same DuckDB type
databricks	`STRING`	`TEXT`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type
databricks	`STRUCT<a: INT>`	`STRUCT(a INT)`	`STRUCT(a INT)`
databricks	`TIMESTAMP`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
databricks	`TIMESTAMP_NTZ`	`TIMESTAMP`	`TIMESTAMP`
databricks	`TINYINT`	`TINYINT`	`TINYINT`
databricks	`VARCHAR(20)`	`TEXT(20)`	`VARCHAR`	TEXT(n) and VARCHAR are the same DuckDB type; DuckDB ignores the length
databricks	`VARIANT`	`VARIANT`	`VARIANT`
databricks	`VOID`	`VARCHAR`	`VARCHAR`
duckdb	`BLOB`	`VARCHAR`	`VARCHAR`
duckdb	`ENUM('a', 'b')`	`ENUM('a', 'b')`	`ENUM('a', 'b')`
duckdb	`HUGEINT`	`INT128`	`INT128`
duckdb	`LIST<INTEGER>`	`INT[]`	`INT[]`
duckdb	`STRUCT(a INTEGER)`	`STRUCT(a INT)`	`STRUCT(a INTEGER)`	DuckDB spelling passes through; INTEGER is INT
duckdb	`TIME WITH TIME ZONE`	`TIMETZ`	`TIME WITH TIME ZONE`	DuckDB spelling passes through; same type as TIMETZ
duckdb	`TIMESTAMP_NS`	`TIMESTAMP_NS`	`TIMESTAMP_NS`
duckdb	`UHUGEINT`	`UINT128`	`UINT128`
duckdb	`UINTEGER`	`UINTEGER`	`UINTEGER`
duckdb	`UNION(num INTEGER, str VARCHAR)`	`UNION(num INT, str TEXT)`	`UNION(num INTEGER, str VARCHAR)`	DuckDB spelling passes through; same types
generic	`DATE`	`DATE`	`DATE`
generic	`DECIMAL(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10,2)`	spacing only
generic	`DECIMAL(38,0)`	`DECIMAL(38, 0)`	`DECIMAL(38,0)`	spacing only
generic	`INTEGER`	`INT`	`INT`
generic	`NUMBER`	`VARCHAR`	`VARCHAR`
generic	`NUMBER(10,2)`	`VARCHAR`	`VARCHAR`
generic	`STANDARD`	`VARCHAR`	`VARCHAR`
generic	`TIMESTAMP`	`TIMESTAMP`	`TIMESTAMP`
generic	`VARCHAR`	`TEXT`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type
generic	`WIDGET`	`VARCHAR`	`VARCHAR`
postgres	`bigint`	`BIGINT`	`BIGINT`
postgres	`bigserial`	`BIGINT`	`BIGINT`
postgres	`bytea`	`VARCHAR`	`VARCHAR`
postgres	`cidr`	`VARCHAR`	`VARCHAR`
postgres	`double precision`	`DOUBLE`	`DOUBLE`
postgres	`geography`	`VARCHAR`	`VARCHAR`
postgres	`geometry`	`VARCHAR`	`VARCHAR`
postgres	`inet`	`INET`	`INET`
postgres	`int4range`	`VARCHAR`	`VARCHAR`
postgres	`integer`	`INT`	`INT`
postgres	`integer[]`	`INT[]`	`INT[]`
postgres	`interval`	`INTERVAL`	`INTERVAL`
postgres	`jsonb`	`JSON`	`JSON`
postgres	`macaddr`	`VARCHAR`	`VARCHAR`
postgres	`macaddr8`	`VARCHAR`	`VARCHAR`
postgres	`money`	`VARCHAR`	`VARCHAR`
postgres	`name`	`VARCHAR`	`VARCHAR`
postgres	`numeric`	`DECIMAL`	`DECIMAL`
postgres	`numeric(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10, 2)`
postgres	`point`	`VARCHAR`	`VARCHAR`
postgres	`real`	`REAL`	`REAL`
postgres	`regclass`	`VARCHAR`	`VARCHAR`
postgres	`serial`	`INT`	`INT`
postgres	`smallint`	`SMALLINT`	`SMALLINT`
postgres	`smallserial`	`SMALLINT`	`SMALLINT`
postgres	`text`	`TEXT`	`TEXT`
postgres	`timestamp with time zone`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
postgres	`tstzrange`	`VARCHAR`	`VARCHAR`
postgres	`uuid`	`UUID`	`UUID`
postgres	`varchar(255)`	`TEXT(255)`	`TEXT(255)`
postgres	`xml`	`VARCHAR`	`VARCHAR`
snowflake	`ARRAY`	`JSON`	`JSON`
snowflake	`BIGINT`	`BIGINT`	`BIGINT`
snowflake	`BINARY`	`VARCHAR`	`VARCHAR`
snowflake	`BOOLEAN`	`BOOLEAN`	`BOOLEAN`
snowflake	`DATE`	`DATE`	`DATE`
snowflake	`DECIMAL(12,4)`	`DECIMAL(12, 4)`	`DECIMAL(12, 4)`
snowflake	`FLOAT`	`DOUBLE`	`DOUBLE`
snowflake	`FLOAT4`	`REAL`	`REAL`
snowflake	`FLOAT8`	`DOUBLE`	`DOUBLE`
snowflake	`GEOGRAPHY`	`VARCHAR`	`VARCHAR`
snowflake	`GEOMETRY`	`VARCHAR`	`VARCHAR`
snowflake	`INT`	`INT`	`INT`
snowflake	`NUMBER`	`DECIMAL(38, 0)`	`DECIMAL(38, 0)`
snowflake	`NUMBER(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10, 2)`
snowflake	`NUMERIC(9,0)`	`DECIMAL(9, 0)`	`DECIMAL(9, 0)`
snowflake	`OBJECT`	`JSON`	`JSON`
snowflake	`REAL`	`REAL`	`REAL`
snowflake	`STRING`	`TEXT`	`TEXT`
snowflake	`TIME`	`TIME`	`TIME`
snowflake	`TIMESTAMP_LTZ`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
snowflake	`TIMESTAMP_LTZ(9)`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
snowflake	`TIMESTAMP_NTZ`	`TIMESTAMP`	`TIMESTAMP`
snowflake	`TIMESTAMP_NTZ(9)`	`TIMESTAMP(9)`	`TIMESTAMP(9)`
snowflake	`TIMESTAMP_TZ`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
snowflake	`VARCHAR`	`TEXT`	`TEXT`
snowflake	`VARCHAR(255)`	`TEXT(255)`	`TEXT(255)`
snowflake	`VARIANT`	`VARIANT`	`JSON`	held as JSON, as SQLQuack's conversion reads it
snowflake	`VECTOR(FLOAT, 3)`	`JSON`	`JSON`

SQL Server has no SQLQuack type parser, so its information_schema names reached the generic fallback, which held datetime2, smalldatetime, datetimeoffset and uniqueidentifier as VARCHAR. A SQL Server pre-mapper now maps them to TIMESTAMP, TIMESTAMPTZ and UUID, bit to BOOLEAN, money and smallmoney to exact decimals, and SQL Server's 8-byte float to DOUBLE. The fallback keeps 4-byte floats (FLOAT, FLOAT4, REAL) as FLOAT instead of widening them to DOUBLE. A panic in the warehouse type parser now surfaces as the ValueError capture already falls back on.

kvlonge · 2026-11-08T04:48:12Z

Before/after type mapping at 7655ef3: every dialect and type pair in the capture tests, plus the SQL Server information_schema names and the float spellings from review. Polyglot on base 97ab62c is compared with this branch. 31 of 156 rows change; the T-SQL datetime2, smalldatetime, datetimeoffset and uniqueidentifier rows keep their timestamp and uuid types, so they don't appear as changes.

Dialect	Warehouse type	Before (Polyglot)	After (SQLQuack)	Change
bigquery	`ARRAY<INT64>`	`BIGINT[]`	`BIGINT[]`
bigquery	`ARRAY<STRUCT<a INT64, b STRING>>`	`STRUCT(a BIGINT, b TEXT)[]`	`STRUCT(a BIGINT, b TEXT)[]`
bigquery	`BIGNUMERIC`	`DECIMAL(38, 5)`	`DECIMAL(38, 5)`
bigquery	`BIGNUMERIC(38,9)`	`DECIMAL(38, 9)`	`DECIMAL(38, 9)`
bigquery	`BOOL`	`BOOLEAN`	`BOOLEAN`
bigquery	`BYTES`	`VARCHAR`	`VARCHAR`
bigquery	`BYTES(10)`	`VARCHAR`	`VARCHAR`
bigquery	`DATE`	`DATE`	`DATE`
bigquery	`DATETIME`	`TIMESTAMP`	`TIMESTAMP`
bigquery	`FLOAT64`	`DOUBLE`	`DOUBLE`
bigquery	`GEOGRAPHY`	`VARCHAR`	`VARCHAR`
bigquery	`INT64`	`BIGINT`	`BIGINT`
bigquery	`INTEGER`	`INT`	`BIGINT`	BigQuery INTEGER is INT64
bigquery	`JSON`	`JSON`	`JSON`
bigquery	`NUMERIC`	`DECIMAL`	`DECIMAL`
bigquery	`NUMERIC(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10, 2)`
bigquery	`RANGE<DATE>`	`VARCHAR`	`VARCHAR`
bigquery	`RANGE<DATETIME>`	`VARCHAR`	`VARCHAR`
bigquery	`RANGE<TIMESTAMP>`	`VARCHAR`	`VARCHAR`
bigquery	`STRING`	`TEXT`	`TEXT`
bigquery	`STRUCT<a INT64>`	`STRUCT(a BIGINT)`	`STRUCT(a BIGINT)`
bigquery	`TIMESTAMP`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
databricks	`ARRAY<FLOAT>`	`FLOAT[]`	`FLOAT[]`
databricks	`ARRAY<INT>`	`INT[]`	`INT[]`
databricks	`DECIMAL(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10,2)`	spacing only
databricks	`DOUBLE`	`DOUBLE`	`DOUBLE`
databricks	`FLOAT`	`FLOAT`	`FLOAT`
databricks	`GEOGRAPHY`	`VARCHAR`	`VARCHAR`
databricks	`GEOMETRY`	`VARCHAR`	`VARCHAR`
databricks	`INTERVAL YEAR`	`INTERVAL YEAR`	`INTERVAL YEAR`
databricks	`MAP<STRING, ARRAY<INT>>`	`MAP(TEXT, INT[])`	`MAP(VARCHAR, INT[])`	TEXT and VARCHAR are the same DuckDB type
databricks	`MAP<STRING, INT>`	`MAP(TEXT, INT)`	`MAP(VARCHAR, INT)`	TEXT and VARCHAR are the same DuckDB type
databricks	`STRING`	`TEXT`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type
databricks	`STRUCT<a: FLOAT>`	`STRUCT(a FLOAT)`	`STRUCT(a FLOAT)`
databricks	`STRUCT<a: INT>`	`STRUCT(a INT)`	`STRUCT(a INT)`
databricks	`TIMESTAMP`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
databricks	`TIMESTAMP_NTZ`	`TIMESTAMP`	`TIMESTAMP`
databricks	`TINYINT`	`TINYINT`	`TINYINT`
databricks	`VARCHAR(20)`	`TEXT(20)`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type; DuckDB ignores the length
databricks	`VARIANT`	`VARIANT`	`VARIANT`
databricks	`VOID`	`VARCHAR`	`VARCHAR`
duckdb	`BLOB`	`VARCHAR`	`VARCHAR`
duckdb	`ENUM('a', 'b')`	`ENUM('a', 'b')`	`ENUM('a', 'b')`
duckdb	`HUGEINT`	`INT128`	`INT128`
duckdb	`LIST<INTEGER>`	`INT[]`	`INT[]`
duckdb	`STRUCT(a INTEGER)`	`STRUCT(a INT)`	`STRUCT(a INTEGER)`	DuckDB spelling passes through unchanged; same type
duckdb	`TIME WITH TIME ZONE`	`TIMETZ`	`TIME WITH TIME ZONE`	DuckDB spelling passes through unchanged; same type
duckdb	`TIMESTAMP_NS`	`TIMESTAMP_NS`	`TIMESTAMP_NS`
duckdb	`UHUGEINT`	`UINT128`	`UINT128`
duckdb	`UINTEGER`	`UINTEGER`	`UINTEGER`
duckdb	`UNION(num INTEGER, str VARCHAR)`	`UNION(num INT, str TEXT)`	`UNION(num INTEGER, str VARCHAR)`	DuckDB spelling passes through unchanged; same type
generic	`DATE`	`DATE`	`DATE`
generic	`DECIMAL(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10,2)`	spacing only
generic	`DECIMAL(38,0)`	`DECIMAL(38, 0)`	`DECIMAL(38,0)`	spacing only
generic	`DOUBLE`	`DOUBLE`	`DOUBLE`
generic	`FLOAT`	`FLOAT`	`FLOAT`
generic	`INTEGER`	`INT`	`INT`
generic	`NUMBER`	`VARCHAR`	`VARCHAR`
generic	`NUMBER(10,2)`	`VARCHAR`	`VARCHAR`
generic	`REAL`	`VARCHAR`	`VARCHAR`
generic	`STANDARD`	`VARCHAR`	`VARCHAR`
generic	`TIMESTAMP`	`TIMESTAMP`	`TIMESTAMP`
generic	`VARCHAR`	`TEXT`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type
generic	`WIDGET`	`VARCHAR`	`VARCHAR`
postgres	`bigint`	`BIGINT`	`BIGINT`
postgres	`bigserial`	`BIGINT`	`BIGINT`
postgres	`bytea`	`VARCHAR`	`VARCHAR`
postgres	`cidr`	`VARCHAR`	`VARCHAR`
postgres	`double precision`	`DOUBLE`	`DOUBLE`
postgres	`geography`	`VARCHAR`	`VARCHAR`
postgres	`geometry`	`VARCHAR`	`VARCHAR`
postgres	`inet`	`INET`	`INET`
postgres	`int4range`	`VARCHAR`	`VARCHAR`
postgres	`integer`	`INT`	`INT`
postgres	`integer[]`	`INT[]`	`INT[]`
postgres	`interval`	`INTERVAL`	`INTERVAL`
postgres	`jsonb`	`JSON`	`JSON`
postgres	`macaddr`	`VARCHAR`	`VARCHAR`
postgres	`macaddr8`	`VARCHAR`	`VARCHAR`
postgres	`money`	`VARCHAR`	`VARCHAR`
postgres	`name`	`VARCHAR`	`VARCHAR`
postgres	`numeric`	`DECIMAL`	`DECIMAL`
postgres	`numeric(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10, 2)`
postgres	`point`	`VARCHAR`	`VARCHAR`
postgres	`real`	`REAL`	`REAL`
postgres	`regclass`	`VARCHAR`	`VARCHAR`
postgres	`serial`	`INT`	`INT`
postgres	`smallint`	`SMALLINT`	`SMALLINT`
postgres	`smallserial`	`SMALLINT`	`SMALLINT`
postgres	`text`	`TEXT`	`TEXT`
postgres	`timestamp with time zone`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
postgres	`tstzrange`	`VARCHAR`	`VARCHAR`
postgres	`uuid`	`UUID`	`UUID`
postgres	`varchar(255)`	`TEXT(255)`	`TEXT(255)`
postgres	`xml`	`VARCHAR`	`VARCHAR`
snowflake	`ARRAY`	`JSON`	`JSON`
snowflake	`BIGINT`	`BIGINT`	`BIGINT`
snowflake	`BINARY`	`VARCHAR`	`VARCHAR`
snowflake	`BOOLEAN`	`BOOLEAN`	`BOOLEAN`
snowflake	`DATE`	`DATE`	`DATE`
snowflake	`DECIMAL(12,4)`	`DECIMAL(12, 4)`	`DECIMAL(12, 4)`
snowflake	`FLOAT`	`DOUBLE`	`DOUBLE`
snowflake	`FLOAT4`	`REAL`	`REAL`
snowflake	`FLOAT8`	`DOUBLE`	`DOUBLE`
snowflake	`GEOGRAPHY`	`VARCHAR`	`VARCHAR`
snowflake	`GEOMETRY`	`VARCHAR`	`VARCHAR`
snowflake	`INT`	`INT`	`INT`
snowflake	`NUMBER`	`DECIMAL(38, 0)`	`DECIMAL(38, 0)`
snowflake	`NUMBER(10,2)`	`DECIMAL(10, 2)`	`DECIMAL(10, 2)`
snowflake	`NUMERIC(9,0)`	`DECIMAL(9, 0)`	`DECIMAL(9, 0)`
snowflake	`OBJECT`	`JSON`	`JSON`
snowflake	`REAL`	`REAL`	`REAL`
snowflake	`STRING`	`TEXT`	`TEXT`
snowflake	`TIME`	`TIME`	`TIME`
snowflake	`TIMESTAMP_LTZ`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
snowflake	`TIMESTAMP_LTZ(9)`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
snowflake	`TIMESTAMP_NTZ`	`TIMESTAMP`	`TIMESTAMP`
snowflake	`TIMESTAMP_NTZ(9)`	`TIMESTAMP(9)`	`TIMESTAMP(9)`
snowflake	`TIMESTAMP_TZ`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
snowflake	`VARCHAR`	`TEXT`	`TEXT`
snowflake	`VARCHAR(255)`	`TEXT(255)`	`TEXT(255)`
snowflake	`VARIANT`	`VARIANT`	`JSON`	held as JSON, as SQLQuack's conversion reads it
snowflake	`VECTOR(FLOAT, 3)`	`JSON`	`JSON`
tsql	`bigint`	`BIGINT`	`BIGINT`
tsql	`binary(16)`	`VARCHAR`	`VARCHAR`
tsql	`bit`	`BIT`	`BOOLEAN`	SQL Server bit is a boolean
tsql	`char(10)`	`TEXT`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type
tsql	`date`	`DATE`	`DATE`
tsql	`datetime`	`TIMESTAMP`	`TIMESTAMP`
tsql	`datetime2`	`TIMESTAMP`	`TIMESTAMP`
tsql	`datetime2(7)`	`TIMESTAMP`	`TIMESTAMP`
tsql	`datetimeoffset`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
tsql	`datetimeoffset(7)`	`TIMESTAMPTZ`	`TIMESTAMPTZ`
tsql	`decimal(18,2)`	`DECIMAL(18, 2)`	`DECIMAL(18,2)`	spacing only
tsql	`float`	`FLOAT`	`DOUBLE`	SQL Server float is 8 bytes
tsql	`int`	`INT`	`INT`
tsql	`money`	`VARCHAR`	`DECIMAL(19,4)`	exact decimal(19,4), as SQL Server stores it
tsql	`nchar(10)`	`TEXT(10)`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type; DuckDB ignores the length
tsql	`ntext`	`TEXT`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type
tsql	`numeric(10,0)`	`DECIMAL(10, 0)`	`DECIMAL(10,0)`	spacing only
tsql	`nvarchar(50)`	`TEXT(50)`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type; DuckDB ignores the length
tsql	`nvarchar(max)`	`TEXT(MAX)`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type; DuckDB ignores the length
tsql	`real`	`REAL`	`FLOAT`	REAL and FLOAT are the same 4-byte DuckDB type
tsql	`smalldatetime`	`TIMESTAMP`	`TIMESTAMP`
tsql	`smallint`	`SMALLINT`	`SMALLINT`
tsql	`smallmoney`	`SMALLMONEY`	`DECIMAL(10,4)`	exact decimal(10,4), as SQL Server stores it
tsql	`sql_variant`	`SQL_VARIANT`	`VARCHAR`	was an invalid DuckDB type
tsql	`text`	`TEXT`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type
tsql	`time`	`TIME`	`TIME`
tsql	`time(7)`	`TIME(7)`	`TIME`	DuckDB TIME takes no precision; was invalid
tsql	`tinyint`	`TINYINT`	`TINYINT`
tsql	`uniqueidentifier`	`UUID`	`UUID`
tsql	`varbinary(max)`	`VARBINARY(MAX)`	`VARCHAR`	binary is held as VARCHAR, as for every dialect; was invalid
tsql	`varchar(50)`	`TEXT(50)`	`VARCHAR`	TEXT and VARCHAR are the same DuckDB type; DuckDB ignores the length
tsql	`varchar(max)`	`VARCHAR(MAX)`	`VARCHAR`	was an invalid DuckDB type
tsql	`xml`	`VARCHAR`	`VARCHAR`

…walks

…line

…d-local-sql # Conflicts: # crates/sqlbuild-analysis/src/lib.rs

SQLQuack's next commit refuses a bare call in a Snowflake FROM clause, which rejects the __ref and __source placeholders that compiled model SQL keeps, so custom-rule syntax trees fail on Snowflake models. This rev keeps the BigQuery parser the warehouse type parsers need.

kvlonge · 2026-11-08T21:00:55Z

Closing: every commit on this branch is already contained in #1374 (the combined parser train), which will land this change. The branch is kept; reopen if #1374 is dropped.

Merge branch 'main' into feat/sqlquack-type-and-local-sql

5ad2d42

kvlonge enabled auto-merge (squash) November 8, 2026 04:57

kvlonge added 3 commits November 8, 2026 05:56

chore(fensu): allow the cast-target walk next to the other JSON tree …

4637fb4

…walks

docs(analysis): keep the warehouse type binding's doc comment to one …

5f253c1

…line

Merge remote-tracking branch 'origin/main' into feat/sqlquack-type-an…

28a28c5

…d-local-sql # Conflicts: # crates/sqlbuild-analysis/src/lib.rs

This was referenced Nov 8, 2026

feat(types): normalize adapter types without Polyglot #1366

Closed

chore(deps): bump sqlquack to 19c51fa2 #1368

Closed

kvlonge disabled auto-merge November 8, 2026 10:03

kvlonge mentioned this pull request Nov 8, 2026

feat: parse rule facts, types and syntax trees with SQLQuack #1374

Closed

kvlonge closed this Nov 8, 2026

Navigation Menu

Repository navigation

feat(scenario): map scenario column types with SQLQuack's type parsers - #1364

feat(scenario): map scenario column types with SQLQuack's type parsers#1364
kvlonge wants to merge 7 commits into
mainfrom
feat/sqlquack-type-and-local-sql

kvlonge commented Nov 8, 2026 •

edited

Loading

Uh oh!

kvlonge commented Nov 8, 2026

Uh oh!

kvlonge commented Nov 8, 2026

Uh oh!

kvlonge commented Nov 8, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

kvlonge commented Nov 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Why

Changes

Verification

Uh oh!

kvlonge commented Nov 8, 2026

Uh oh!

kvlonge commented Nov 8, 2026

Uh oh!

kvlonge commented Nov 8, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

kvlonge commented Nov 8, 2026 •

edited

Loading