fix: adds bounds checking because pandas now handles microsecond reso… (#166)

* fix: adds bounds checking because pandas now handles microsecond resolution

* Update db_dtypes/__init__.py

Co-authored-by: Tim Swast <[email protected]>

Co-authored-by: Tim Swast <[email protected]>

Author: chalmerlowe

db_dtypes/__init__.py

@@ -23,9 +23,11 @@
 import packaging.version
 import pandas
 import pandas.api.extensions
+from pandas.errors import OutOfBoundsDatetime
 import pyarrow
 import pyarrow.compute
 
+
 from db_dtypes.version import __version__
 from db_dtypes import core
 
@@ -143,6 +145,7 @@ def _datetime(
             second = parsed.group("seconds")
             fraction = parsed.group("fraction")
             nanosecond = int(fraction.ljust(9, "0")[:9]) if fraction else 0
+
             return pandas.Timestamp(
                 year=1970,
                 month=1,
@@ -263,7 +266,17 @@ def _datetime(
             year = int(match.group("year"))
             month = int(match.group("month"))
             day = int(match.group("day"))
-            return pandas.Timestamp(year=year, month=month, day=day).to_datetime64()
+
+            dateObj = pandas.Timestamp(
+                year=year,
+                month=month,
+                day=day,
+            )
+            if pandas.Timestamp.min < dateObj < pandas.Timestamp.max:
+                return dateObj.to_datetime64()
+            else:  # pragma: NO COVER
+                # TODO(#166): Include these lines in coverage when pandas 2.0 is released.
+                raise OutOfBoundsDatetime("Out of bounds", scalar)  # pragma: NO COVER
         else:
             raise TypeError("Invalid value type", scalar)
 

tests/unit/test_date.py

@@ -18,6 +18,7 @@
 import numpy
 import numpy.testing
 import pandas
+from pandas.errors import OutOfBoundsDatetime
 import pandas.testing
 import pytest
 
@@ -143,15 +144,25 @@ def test_date_set_slice_null():
         ("2021-2-99", "day is out of range for month"),
         ("2021-99-1", "month must be in 1[.][.]12"),
         ("10000-1-1", "year 10000 is out of range"),
-        # Outside of min/max values pandas.Timestamp.
+    ],
+)
+def test_date_parsing_errors(value, error):
+    with pytest.raises(ValueError, match=error):
+        pandas.Series([value], dtype="dbdate")
+
+
+@pytest.mark.parametrize(
+    "value, error",
+    [
+        # Values that are outside of the min/max values allowed by pandas.Timestamp
         ("0001-01-01", "Out of bounds"),
         ("9999-12-31", "Out of bounds"),
         ("1677-09-21", "Out of bounds"),
         ("2262-04-12", "Out of bounds"),
     ],
 )
-def test_date_parsing_errors(value, error):
-    with pytest.raises(ValueError, match=error):
+def test_date_parsing_errors_out_of_bounds(value, error):
+    with pytest.raises(OutOfBoundsDatetime, match=error):
         pandas.Series([value], dtype="dbdate")