From 19b08a8b50c46c0cb10105759b95bc359154ed98 Mon Sep 17 00:00:00 2001
From: adenzhou1350 <209601943+adenzhou1350@users.noreply.github.com>
Date: Tue, 6 Oct 2026 02:04:40 +0800
Subject: [PATCH] Recognize JOIN before parenthesized subqueries

---
 CHANGELOG                      |  3 ++-
 sqlparse/keywords.py           |  5 +++--
 tests/test_join_parenthesis.py | 32 ++++++++++++++++++++++++++++++++
 3 files changed, 37 insertions(+), 3 deletions(-)
 create mode 100644 tests/test_join_parenthesis.py

diff --git a/CHANGELOG b/CHANGELOG
index 44d5938e..06d5689e 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -1,7 +1,8 @@
 Development Version
 -------------------
 
-Nothing yet.
+* Recognize JOIN immediately before a parenthesized subquery as a keyword
+  instead of a function name (issue615).
 
 
 Release 0.6.0 (Aug 13, 2026)
diff --git a/sqlparse/keywords.py b/sqlparse/keywords.py
index dd6e5d15..20361cd7 100644
--- a/sqlparse/keywords.py
+++ b/sqlparse/keywords.py
@@ -157,6 +157,9 @@ def find_delimited_spans(text):
     # FIXME(atronah): never match,
     # because `re.match` doesn't work with look-behind regexp feature
     (r'(?<=\.)[A-ZÀ-Ü]\w*', tokens.Name),  # .'Name'
+    # JOIN before a parenthesis is a clause, not a function (issue615).
+    (r'((LEFT\s+|RIGHT\s+|FULL\s+)?(INNER\s+|OUTER\s+|STRAIGHT\s+)?'
+     r'|(CROSS\s+|NATURAL\s+)?)?JOIN\b', tokens.Keyword),
     (r'[A-ZÀ-Ü]\w*(?=\()', tokens.Name),  # side effect: change kw to func
     (r'-?0x[\dA-F]+', tokens.Number.Hexadecimal),
     (r'-?\d+(\.\d+)?E-?\d+', tokens.Number.Float),
@@ -171,8 +174,6 @@ def find_delimited_spans(text):
     # cannot be preceded by word character or a right bracket --
     # otherwise it's probably an array index
     (r'(?<![\w\])])(\[[^\]\[]+\])', tokens.Name),
-    (r'((LEFT\s+|RIGHT\s+|FULL\s+)?(INNER\s+|OUTER\s+|STRAIGHT\s+)?'
-     r'|(CROSS\s+|NATURAL\s+)?)?JOIN\b', tokens.Keyword),
     (r'END(\s+IF|\s+LOOP|\s+WHILE|\s+FOR|\s+CASE)?\b', tokens.Keyword),
     (r'IF\s+(NOT\s+)?EXISTS\b', tokens.Keyword),
     (r'NOT\s+NULL\b', tokens.Keyword),
diff --git a/tests/test_join_parenthesis.py b/tests/test_join_parenthesis.py
new file mode 100644
index 00000000..951536e2
--- /dev/null
+++ b/tests/test_join_parenthesis.py
@@ -0,0 +1,32 @@
+import pytest
+
+import sqlparse
+from sqlparse import sql
+from sqlparse import tokens as T
+from sqlparse.lexer import tokenize
+
+
+@pytest.mark.parametrize('join', [
+    'JOIN', 'join', 'LEFT JOIN', 'LEFT OUTER JOIN', 'CROSS JOIN', 'NATURAL JOIN',
+])
+def test_join_before_parenthesized_subquery(join):
+    condition = '' if join in ('CROSS JOIN', 'NATURAL JOIN') else ' on child.parent_id=p.id'
+    query = f'select * from child {join}(select * from parent) as p{condition};'
+    statement = sqlparse.parse(query)[0]
+    significant = [token for token in statement if not token.is_whitespace]
+
+    assert significant[4].match(T.Keyword, join.upper())
+    assert isinstance(significant[3], sql.Identifier)
+    assert significant[3].get_real_name() == 'child'
+    assert isinstance(significant[5], sql.Identifier)
+    assert significant[5].get_alias() == 'p'
+    assert str(statement) == query
+
+
+@pytest.mark.parametrize('name', ['joined', 'join_table', 'schema.join', 'schema.JOIN'])
+def test_join_function_name_controls(name):
+    query = f'{name}(1)'
+    names = [value for ttype, value in tokenize(query) if ttype is T.Name]
+    assert names == name.split('.')
+    expected_group = sql.Identifier if '.' in name else sql.Function
+    assert isinstance(sqlparse.parse(query)[0].tokens[0], expected_group)