From 237662937ddff2e6e4e99ee3107e393e7d9b6af8 Mon Sep 17 00:00:00 2001
From: Rahul-cs50 <rahulseervi.dev@gmail.com>
Date: Sat, 31 Oct 2026 14:14:31 +0530
Subject: [PATCH] Read a naive datetime in older_than as local time

expire_snapshots().older_than() converted its cutoff with
datetime_to_millis, which reads a naive datetime as UTC. The docs call
it with datetime.now() - timedelta(...), a naive local time, so east of
UTC the cutoff landed later and expired snapshots that should be kept,
and west of UTC it landed earlier. A naive datetime is now read as local
time, as datetime.timestamp() does; an aware one is unchanged.

Closes #4182
---
 pyiceberg/table/update/snapshot.py   |  3 +++
 tests/table/test_expire_snapshots.py | 29 ++++++++++++++++++++++++++++
 2 files changed, 32 insertions(+)

diff --git a/pyiceberg/table/update/snapshot.py b/pyiceberg/table/update/snapshot.py
index 8feebb9901..d0a5ed476d 100644
--- a/pyiceberg/table/update/snapshot.py
+++ b/pyiceberg/table/update/snapshot.py
@@ -1418,11 +1418,14 @@ def older_than(self, dt: datetime) -> ExpireSnapshots:
 
         Args:
             dt (datetime): Only snapshots with datetime < this value will be expired.
+                A naive datetime is read as local time, like datetime.now() returns it.
 
         Returns:
             This for method chaining.
         """
         protected_ids = self._get_protected_snapshot_ids()
+        if dt.tzinfo is None:
+            dt = dt.astimezone()
         expire_from = datetime_to_millis(dt)
         for snapshot in self._transaction.table_metadata.snapshots:
             if snapshot.timestamp_ms < expire_from and snapshot.snapshot_id not in protected_ids:
diff --git a/tests/table/test_expire_snapshots.py b/tests/table/test_expire_snapshots.py
index 106e5b786c..a393b84b46 100644
--- a/tests/table/test_expire_snapshots.py
+++ b/tests/table/test_expire_snapshots.py
@@ -15,6 +15,7 @@
 # specific language governing permissions and limitations
 # under the License.
 import threading
+import time
 from datetime import datetime, timedelta
 from unittest.mock import MagicMock, Mock
 from uuid import uuid4
@@ -174,6 +175,34 @@ def test_expire_snapshots_by_timestamp_skips_protected(table_v2: Table) -> None:
     assert remove_update.snapshot_ids == []
 
 
+def test_expire_snapshots_older_than_reads_naive_datetime_as_local_time(table_v2: Table, monkeypatch: pytest.MonkeyPatch) -> None:
+    """Test that older_than reads a naive datetime, as datetime.now() returns it, in the local time zone."""
+    if not hasattr(time, "tzset"):
+        pytest.skip("time.tzset is not available on Windows")
+    EXPIRE_SNAPSHOT = 3051729675574597004
+    KEEP_SNAPSHOT = 3055729675574597004
+    expire_snapshot_ms = 1515100955770
+
+    table_v2.metadata = table_v2.metadata.model_copy(
+        update={"refs": {"main": MagicMock(snapshot_id=KEEP_SNAPSHOT, snapshot_ref_type="branch")}}
+    )
+
+    # A zone east of UTC, so reading local time as UTC moves the cutoff nine hours earlier
+    monkeypatch.setenv("TZ", "Asia/Tokyo")
+    time.tzset()
+    try:
+        local_snapshot_time = datetime.fromtimestamp(expire_snapshot_ms / 1000)
+        for cutoff, expected in (
+            (local_snapshot_time - timedelta(minutes=1), set()),
+            (local_snapshot_time + timedelta(minutes=1), {EXPIRE_SNAPSHOT}),
+        ):
+            expire = table_v2.maintenance.expire_snapshots().older_than(cutoff)
+            assert expire._snapshot_ids_to_expire == expected
+    finally:
+        monkeypatch.undo()
+        time.tzset()
+
+
 def test_expire_snapshots_by_ids(table_v2: Table) -> None:
     """Test that multiple unprotected snapshots can be expired by IDs."""
     EXPIRE_SNAPSHOT_1 = 3051729675574597004