fix: Workaround for dates < 1970 on Windows

tests/unit/ibis_addon/test_operations.py

@@ -22,6 +22,8 @@
 CLIENT = ibis.pandas.connect({"table": TABLE_DF})
 WHERE_FILTER = "id > 100"
 
+SECONDS_IN_A_DAY = 60 * 60 * 24
+
 
 @pytest.fixture
 def module_under_test():
@@ -45,3 +47,37 @@ def test_format_raw_sql_expr(module_under_test):
     raw_sql = operations.format_raw_sql(ibis_table.column, raw_sql_column_expr)
 
     assert raw_sql == WHERE_FILTER
+
+
+@pytest.mark.parametrize(
+    "test_input,expected",
+    [
+        (
+            "1970-01-01",
+            0,
+        ),
+        (
+            "1970-01-01 00:00:01",
+            1,
+        ),
+        (
+            "1970-01-02",
+            SECONDS_IN_A_DAY,
+        ),
+        (
+            "1970-02-01 00:00:01",
+            (SECONDS_IN_A_DAY * 31) + 1,
+        ),
+        (
+            "1969-12-31",
+            -SECONDS_IN_A_DAY,
+        ),
+        (
+            "1969-12-31 23:59:00",
+            -60,
+        ),
+    ],
+)
+def test_string_to_epoch(module_under_test, test_input: str, expected: int):
+    result = module_under_test.string_to_epoch(test_input)
+    assert result == expected

third_party/ibis/ibis_addon/operations.py

@@ -23,6 +23,7 @@
 non-textual languages.
 """
 import datetime
+import dateutil
 
 import google.cloud.bigquery as bq
 import ibis
@@ -588,19 +589,18 @@ def _bigquery_field_to_ibis_dtype(field):
     return ibis_type
 
 
-def string_to_epoch(ts: str):
-    "Function to convert string timestamp to epoch seconds"
-    from dateutil.parser import isoparse
-    from dateutil.tz import UTC
-    from datetime import datetime, timezone
-
+def string_to_epoch(ts: str) -> int:
+    """Function to convert string timestamp to epoch seconds"""
     try:
-        parsed_ts = isoparse(ts).astimezone(UTC)
-        return (parsed_ts - datetime(1970, 1, 1, tzinfo=timezone.utc)).total_seconds()
-    except ValueError:
+        parsed_ts = dateutil.parser.isoparse(ts).astimezone(dateutil.tz.UTC)
+        return (
+            parsed_ts - datetime.datetime(1970, 1, 1, tzinfo=datetime.timezone.utc)
+        ).total_seconds()
+    except (ValueError, OSError):
         # Support DATE '0001-01-01' which throws error when converted to UTC
-        parsed_ts = isoparse(ts)
-        return (parsed_ts - datetime(1970, 1, 1)).total_seconds()
+        # Catching OSError above because all dates prior to 1970 fail in astimezone on Windows.
+        parsed_ts = dateutil.parser.isoparse(ts)
+        return (parsed_ts - datetime.datetime(1970, 1, 1)).total_seconds()
 
 
 @execute_node.register(ops.ExtractEpochSeconds, (datetime.datetime, pd.Series))