From 6a4ea3dfaefcfefdd9682328a72630be3f66b140 Mon Sep 17 00:00:00 2001
From: adenzhou1350 <209601943+adenzhou1350@users.noreply.github.com>
Date: Tue, 6 Oct 2026 14:38:35 +0800
Subject: [PATCH] Group comparisons across intervening comments

---
 sqlparse/engine/grouping.py | 14 +++++++----
 tests/test_format.py        | 14 +++++++++++
 tests/test_grouping.py      | 49 +++++++++++++++++++++++++++++++++++++
 3 files changed, 72 insertions(+), 5 deletions(-)

diff --git a/sqlparse/engine/grouping.py b/sqlparse/engine/grouping.py
index d8cfa9e1..43cc7c6f 100644
--- a/sqlparse/engine/grouping.py
+++ b/sqlparse/engine/grouping.py
@@ -240,8 +240,10 @@ def post(tlist, pidx, tidx, nidx):
         return pidx, nidx
 
     valid_prev = valid_next = valid
+    # Comments between an operator and operand belong to the comparison,
+    # not a separate identifier list (issue604).
     _group(tlist, sql.Comparison, match,
-           valid_prev, valid_next, post, extend=False)
+           valid_prev, valid_next, post, extend=False, skip_cm=True)
 
 
 @recurse(sql.Identifier)
@@ -481,7 +483,8 @@ def _group(tlist, cls, match,
            post=None,
            extend=True,
            recurse=True,
-           depth=0
+           depth=0,
+           skip_cm=False
            ):
     """Groups together tokens that are joined by a middle token. i.e. x < y"""
     if MAX_GROUPING_DEPTH is not None and depth > MAX_GROUPING_DEPTH:
@@ -505,15 +508,16 @@ def _group(tlist, cls, match,
         if tidx < 0:  # tidx shouldn't get negative
             continue
 
-        if token.is_whitespace:
+        if token.is_whitespace or (skip_cm and imt(token, i=sql.Comment,
+                                                  t=T.Comment)):
             continue
 
         if recurse and token.is_group and not isinstance(token, cls):
             _group(token, cls, match, valid_prev, valid_next,
-                   post, extend, True, depth + 1)
+                   post, extend, True, depth + 1, skip_cm)
 
         if match(token):
-            nidx, next_ = tlist.token_next(tidx)
+            nidx, next_ = tlist.token_next(tidx, skip_cm=skip_cm)
             if prev_ and valid_prev(prev_) and valid_next(next_):
                 from_idx, to_idx = post(tlist, pidx, tidx, nidx)
                 grp = tlist.group_tokens(cls, from_idx, to_idx, extend=extend)
diff --git a/tests/test_format.py b/tests/test_format.py
index 93495067..2efac1f2 100644
--- a/tests/test_format.py
+++ b/tests/test_format.py
@@ -73,6 +73,20 @@ def test_strip_comments_invalid_option(self):
         with pytest.raises(SQLParseError):
             sqlparse.format(sql, strip_comments=None)
 
+    @pytest.mark.parametrize('comment', (
+        '/* comment */', '/* first line\nsecond line */', '-- comment\n',
+    ))
+    def test_reindent_commented_comparison_columns(self, comment):
+        # issue604: an ungrouped comparison split one SELECT list into
+        # nested lists, so every comment increased subsequent indentation.
+        text = f'SELECT leadcol = leadvalue, aaa = {comment} bbb, ccc = ccc, ddd = ddd FROM t'
+        result = sqlparse.format(text, strip_comments=True, reindent=True)
+        assert result == ('SELECT leadcol = leadvalue,\n'
+                          '       aaa = bbb,\n'
+                          '       ccc = ccc,\n'
+                          '       ddd = ddd\n'
+                          'FROM t')
+
     def test_strip_comments_multi(self):
         sql = '/* sql starts here */\nselect'
         res = sqlparse.format(sql, strip_comments=True)
diff --git a/tests/test_grouping.py b/tests/test_grouping.py
index 20fab9b7..cdfb2e1c 100644
--- a/tests/test_grouping.py
+++ b/tests/test_grouping.py
@@ -488,6 +488,55 @@ def test_comparison_with_keywords():
     assert isinstance(p.tokens[0], sql.Comparison)
 
 
+@pytest.mark.parametrize('left_comment,right_comment', (
+    ('/* before */', ''),
+    ('', '/* after */'),
+    ('/* before */', '/* after */'),
+    ('-- before\n', '-- after\n'),
+    ('/* first */ /* second */', '/* third */'),
+    ('/*+ hint */', '/*+ hint */'),
+))
+def test_comparison_with_comments(left_comment, right_comment):
+    text = f'foo {left_comment} = {right_comment} bar'
+    statement = sqlparse.parse(text)[0]
+    assert len(statement.tokens) == 1
+    comparison = statement.tokens[0]
+    assert isinstance(comparison, sql.Comparison)
+    assert comparison.left.get_real_name() == 'foo'
+    assert comparison.right.value == 'bar'
+    assert str(comparison) == text
+    assert any(token.ttype in T.Comment for token in comparison.flatten())
+
+
+def test_comparison_with_comments_in_parenthesis():
+    statement = sqlparse.parse('(foo /* before */ = /* after */ bar)')[0]
+    comparison = statement.tokens[0].tokens[1]
+    assert isinstance(comparison, sql.Comparison)
+    assert comparison.left.get_real_name() == 'foo'
+    assert comparison.right.value == 'bar'
+
+
+@pytest.mark.parametrize('text,right', (
+    ('foo >= /* comment */ 25.5', '25.5'),
+    ("foo NOT LIKE /* comment */ 'bar'", "'bar'"),
+    ('foo = /* comment */ DATE(bar)', 'DATE(bar)'),
+    ('foo = /* comment */ NULL', 'NULL'),
+))
+def test_comparison_with_commented_operands(text, right):
+    statement = sqlparse.parse(text)[0]
+    comparison = statement.tokens[0]
+    assert isinstance(comparison, sql.Comparison)
+    assert comparison.right.value == right
+    assert str(comparison) == text
+
+
+def test_comparison_does_not_cross_statement_after_comment():
+    text = 'foo = /* comment */; SELECT bar'
+    first, second = sqlparse.parse(text)
+    assert not any(isinstance(token, sql.Comparison) for token in first.tokens)
+    assert str(first) + str(second) == text
+
+
 def test_comparison_with_floats():
     # issue145
     p = sqlparse.parse('foo = 25.5')[0]