BigQuery: explicitly define data types (#7305)

Co-authored-by: Annebelle Olminkhof <20630578+annebelleo@users.noreply.github.com>

src/sqlfluff/dialects/dialect_bigquery.py

@@ -290,6 +290,40 @@ bigquery_dialect.replace(
             casefold=str.upper,
         )
     ),
+    DatatypeIdentifierSegment=SegmentGenerator(
+        lambda dialect: MultiStringParser(
+            [
+                "INT64",
+                "INT",
+                "SMALLINT",
+                "INTEGER",
+                "BIGINT",
+                "TINYINT",
+                "BYTEINT",
+                "FLOAT64",
+                "NUMERIC",
+                "DECIMAL",
+                "BIGNUMERIC",
+                "BIGDECIMAL",
+                "BOOL",
+                "BOOLEAN",
+                "STRING",
+                "BYTES",
+                "DATE",
+                "DATETIME",
+                "TIME",
+                "TIMESTAMP",
+                "GEOGRAPHY",
+                "INTERVAL",
+                "JSON",
+                "RANGE",
+                "ARRAY",
+                "STRUCT",
+            ],
+            CodeSegment,
+            type="data_type_identifier",
+        )
+    ),
     FunctionContentsExpressionGrammar=OneOf(
         Ref("DatetimeUnitSegment"),
         Ref("DatePartWeekSegment"),

test/dialects/bigquery_test.py

@@ -4,7 +4,7 @@ import hypothesis.strategies as st
 import pytest
 from hypothesis import example, given, note, settings
 
-from sqlfluff.core import FluffConfig
+from sqlfluff.core import FluffConfig, Linter
 from sqlfluff.core.parser import Lexer, Parser
 
 
@@ -90,3 +90,11 @@ def test_bigquery_table_reference_segment_iter_raw_references(
             orp.part for orp in table_reference.iter_raw_references()
         ]
         assert reference_parts == actual_reference_parts
+
+
+def test_cast_as_float_fails():
+    """CAST(... AS FLOAT) should fail for BigQuery (only FLOAT64 allowed)."""
+    sql = "SELECT CAST('4.0' AS FLOAT)"
+    parsed = Linter(dialect="bigquery").parse_string(sql)
+    # Parsing produces a parse error for the unsupported `FLOAT` type.
+    assert parsed.violations

test/fixtures/dialects/bigquery/create_js_function_complex_types.sql

@@ -4,7 +4,7 @@ qs(
     foo2 ARRAY<STRING>,
     foo3 STRUCT<x INT64>,
     foo4 STRUCT<x INT64, y INT64>,
-    foo5 STRUCT<a ARRAY<FLOAT>, b STRUCT<x INT64, y INT64>>
+    foo5 STRUCT<a ARRAY<FLOAT64>, b STRUCT<x INT64, y INT64>>
 )
 RETURNS STRUCT<product_id ARRAY<INT64>>
 LANGUAGE js AS """

test/fixtures/dialects/bigquery/create_js_function_complex_types.yml

@@ -3,7 +3,7 @@
 # computed by SQLFluff when running the tests. Please run
 # `python test/generate_parse_fixture_yml.py`  to generate them after adding or
 # altering SQL files.
-_hash: d1fa0d72cc2af0530c5f21c4cea3da882175a110632f386d1752c6136a1bb15f
+_hash: 09fea1147a6721122baff61b940cabd21edd638edb25e3829479fea628e1b460
 file:
   statement:
     create_function_statement:
@@ -66,7 +66,7 @@ file:
                     keyword: ARRAY
                     start_angle_bracket: <
                     data_type:
-                      data_type_identifier: FLOAT
+                      data_type_identifier: FLOAT64
                     end_angle_bracket: '>'
               - comma: ','
               - parameter: b

test/fixtures/dialects/bigquery/select_with_cast.yml

@@ -3,7 +3,7 @@
 # computed by SQLFluff when running the tests. Please run
 # `python test/generate_parse_fixture_yml.py`  to generate them after adding or
 # altering SQL files.
-_hash: b3eb84334d4a93f639f727dd59f0833c1332eab98414140b0076dae106703690
+_hash: bfa63bceb292ee595dc0f2d449ee416a35c43b97ae6289ac6d6844c8bfdd3d3b
 file:
 - statement:
     select_statement:
@@ -87,8 +87,9 @@ file:
               - data_type:
                   data_type_identifier: STRING
               - expression:
-                  data_type:
-                    data_type_identifier: FORMAT
+                  column_reference:
+                    naked_identifier: FORMAT
+              - expression:
                   quoted_literal: "'ASCII'"
               - end_bracket: )
           alias_expression:
@@ -114,8 +115,9 @@ file:
               - data_type:
                   data_type_identifier: STRING
               - expression:
-                  data_type:
-                    data_type_identifier: FORMAT
+                  column_reference:
+                    naked_identifier: FORMAT
+              - expression:
                   quoted_literal: "'YYYY-MM-DD HH24:MI:SS TZH:TZM'"
               - end_bracket: )
           alias_expression: