BigQuery: Fix parsing parameterized data types (#3735)

yoichi · barrywhart · web-flow · commit bc147334a8a8 · 2022-08-14T10:02:00.000-04:00
* BigQuery: Fix parsing parameterized data types

* Replace the type in a comment with a supported one

* Make the param[, ...] part more specific

* Extend DatatypeSegment to support parameterized data types

Co-authored-by: Barry Hart &lt;barrywhart@yahoo.com&gt;
diff --git a/src/sqlfluff/dialects/dialect_bigquery.py b/src/sqlfluff/dialects/dialect_bigquery.py
@@ -960,7 +960,11 @@ class DatatypeSegment(ansi.DatatypeSegment):
     """
 
     match_grammar = OneOf(  # Parameter type
-        Ref("DatatypeIdentifierSegment"),  # Simple type
+        Sequence(
+            Ref("DatatypeIdentifierSegment"),  # Simple type
+            # https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#parameterized_data_types
+            Bracketed(Delimited(Ref("NumericLiteralSegment")), optional=True),
+        ),
         Sequence("ANY", "TYPE"),  # SQL UDFs can specify this "type"
         Ref("SimpleArrayTypeGrammar"),
         Ref("StructTypeSegment"),
@@ -1372,7 +1376,6 @@ class ColumnDefinitionSegment(ansi.ColumnDefinitionSegment):
     match_grammar: Matchable = Sequence(
         Ref("SingleIdentifierGrammar"),  # Column name
         Ref("DatatypeSegment"),  # Column type
-        Bracketed(Anything(), optional=True),  # For types like VARCHAR(100)
         AnyNumberOf(
             Ref("ColumnConstraintSegment", optional=True),
         ),
diff --git a/test/fixtures/dialects/bigquery/declare_variable.sql b/test/fixtures/dialects/bigquery/declare_variable.sql
@@ -2,14 +2,18 @@ declare var1 int64;
 declare var2, var3 string;
 declare var4 default 'value';
 declare var5 int64 default 1 + 2;
+declare var6 string(10);
+declare var7 numeric(5, 2);
 declare arr1 array<string>;
 declare arr2 default ['one', 'two'];
 declare arr3 default [];
 declare arr4 array<string> default ['one', 'two'];
+declare arr5 array<string(10)>;
 declare str1 struct<f1 string, f2 string>;
 declare str2 struct<f1 string, f2 string> default struct('one', 'two');
 declare str3 default struct('one', 'two');
 declare str4 struct<f1 string, f2 string> default ('one', 'two');
+declare str5 struct<f1 string(10), f2 string(10)>;
 -- Defining variables in quoted names
 declare `var1` string;
 declare `var1` string default 'value';
diff --git a/test/fixtures/dialects/bigquery/declare_variable.yml b/test/fixtures/dialects/bigquery/declare_variable.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: e7c3d1f233e5041e5097f81128665af7895b563ccfcf6e247911f714db0b0c04
+_hash: a7344a57c0e59ce697ccfabc4e78ada5ea7a5f53123ec3a8839bca556863d42d
 file:
 - statement:
     declare_segment:
@@ -40,6 +40,30 @@ file:
       - binary_operator: +
       - numeric_literal: '2'
 - statement_terminator: ;
+- statement:
+    declare_segment:
+      keyword: declare
+      naked_identifier: var6
+      data_type:
+        data_type_identifier: string
+        bracketed:
+          start_bracket: (
+          numeric_literal: '10'
+          end_bracket: )
+- statement_terminator: ;
+- statement:
+    declare_segment:
+      keyword: declare
+      naked_identifier: var7
+      data_type:
+        data_type_identifier: numeric
+        bracketed:
+        - start_bracket: (
+        - numeric_literal: '5'
+        - comma: ','
+        - numeric_literal: '2'
+        - end_bracket: )
+- statement_terminator: ;
 - statement:
     declare_segment:
       keyword: declare
@@ -90,6 +114,21 @@ file:
       - quoted_literal: "'two'"
       - end_square_bracket: ']'
 - statement_terminator: ;
+- statement:
+    declare_segment:
+      keyword: declare
+      naked_identifier: arr5
+      data_type:
+        keyword: array
+        start_angle_bracket: <
+        data_type:
+          data_type_identifier: string
+          bracketed:
+            start_bracket: (
+            numeric_literal: '10'
+            end_bracket: )
+        end_angle_bracket: '>'
+- statement_terminator: ;
 - statement:
     declare_segment:
       keyword: declare
@@ -174,6 +213,31 @@ file:
         - quoted_literal: "'two'"
         - end_bracket: )
 - statement_terminator: ;
+- statement:
+    declare_segment:
+      keyword: declare
+      naked_identifier: str5
+      data_type:
+        struct_type:
+        - keyword: struct
+        - start_angle_bracket: <
+        - parameter: f1
+        - data_type:
+            data_type_identifier: string
+            bracketed:
+              start_bracket: (
+              numeric_literal: '10'
+              end_bracket: )
+        - comma: ','
+        - parameter: f2
+        - data_type:
+            data_type_identifier: string
+            bracketed:
+              start_bracket: (
+              numeric_literal: '10'
+              end_bracket: )
+        - end_angle_bracket: '>'
+- statement_terminator: ;
 - statement:
     declare_segment:
       keyword: declare
