Fix #377 #390 #394

sql_metadata/generalizator.py

@@ -47,7 +47,8 @@ def without_comments(self) -> str:
         :rtype: str
         """
         sql = sqlparse.format(self._raw_query, strip_comments=True)
-        sql = re.sub(r"\s{2,}", " ", sql)
+        sql = sql.replace("\n", " ")
+        sql = re.sub(r"[ \t]+", " ", sql)
         return sql
 
     @property

sql_metadata/parser.py

@@ -533,12 +533,16 @@ def subqueries(self) -> Dict:
                 ):
                     current_subquery.append(inner_token)
                     inner_token = inner_token.next_token
+
+                query_name = None
                 if inner_token.next_token.value in self.subqueries_names:
                     query_name = inner_token.next_token.value
-                else:
+                elif inner_token.next_token.is_as_keyword:
                     query_name = inner_token.next_token.next_token.value
+
                 subquery_text = "".join([x.stringified_token for x in current_subquery])
-                subqueries[query_name] = subquery_text
+                if query_name is not None:
+                    subqueries[query_name] = subquery_text
 
             token = token.next_token
 
@@ -622,7 +626,7 @@ def without_comments(self) -> str:
         """
         Removes comments from SQL query
         """
-        return Generalizator(self.query).without_comments
+        return Generalizator(self._raw_query).without_comments
 
     @property
     def generalize(self) -> str:
@@ -865,7 +869,7 @@ def _determine_opening_parenthesis_type(self, token: SQLToken):
             # we are in columns and in a column subquery definition
             token.is_column_definition_start = True
         elif (
-            token.previous_token.is_as_keyword
+            token.previous_token_not_comment.is_as_keyword
             and token.last_keyword_normalized != "WINDOW"
         ):
             # window clause also contains AS keyword, but it is not a query

sql_metadata/token.py

@@ -369,6 +369,15 @@ def next_token_not_comment(self):
             return self.next_token.next_token_not_comment
         return self.next_token
 
+    @property
+    def previous_token_not_comment(self):
+        """
+        Property returning previous non-comment token
+        """
+        if self.previous_token and self.previous_token.is_comment:
+            return self.previous_token.previous_token_not_comment
+        return self.previous_token
+
     def is_constraint_definition_inside_create_table_clause(
         self, query_type: str
     ) -> bool:

test/test_comments.py

@@ -205,3 +205,11 @@ def test_next_token_not_comment_on_non_comments():
         select_tok.next_token.next_token
         == select_tok.next_token_not_comment.next_token_not_comment
     )
+
+
+def test_without_comments_for_multiline_query():
+    query = """SELECT * -- comment
+        FROM table
+        WHERE table.id = '123'"""
+    parser = Parser(query)
+    assert parser.without_comments == """SELECT * FROM table WHERE table.id = '123'"""

test/test_getting_columns.py

@@ -264,6 +264,14 @@ def test_columns_with_comments():
         "order_by": ["cl_sortkey"],
     }
 
+    parser = Parser(
+        """WITH aa AS --sdfsdfsdf 
+        (SELECT C1, C2 FROM T1) 
+        SELECT C1, C2 FROM aa"""
+    )
+    assert parser.columns == ["C1", "C2"]
+    assert parser.columns_dict == {"select": ["C1", "C2"]}
+
 
 def test_columns_with_keyword_aliases():
     parser = Parser(
@@ -477,3 +485,37 @@ def test_having_columns():
         "group_by": ["Country"],
         "having": ["CustomerID"],
     }
+
+
+def test_nested_queries():
+    query = """
+    SELECT max(dt) FROM
+        (
+         SELECT max(dt) as dt FROM t      
+      UNION ALL
+          SELECT max(dt) as dt FROM t2
+        )
+    """
+    parser = Parser(query)
+    assert parser.columns == ["dt"]
+    assert parser.columns_dict == {"select": ["dt"]}
+
+    query = """
+    SELECT max(dt) FROM
+        (
+         SELECT max(dt) as dt FROM t      
+        )
+    """
+    parser = Parser(query)
+    assert parser.columns == ["dt"]
+    assert parser.columns_dict == {"select": ["dt"]}
+
+    query = """
+    SELECT max(dt) FROM
+        (
+         SELECT dt FROM t      
+        )
+    """
+    parser = Parser(query)
+    assert parser.columns == ["dt"]
+    assert parser.columns_dict == {"select": ["dt"]}