Fix #775: handle empty datasets in get_type with informative error

src/hdmf/validate/validator.py

@@ -175,6 +175,11 @@ def _get_type_from_dtype_attr(data: Any, builder_dtype: list | None) -> tuple[st
 
 def get_type(data, builder_dtype=None):
     """Return a tuple of (the string representation of the type, the format of the string data) for the given data."""
+    if data is None:
+        return None, None
+    # Check for empty data safely
+    if hasattr(data, "__len__") and getattr(data, 'shape', (1,)) != () and len(data) == 0:
+        raise EmptyArrayError("Dataset is empty; cannot determine type.")
     # String data
     if isinstance(data, str):
         return 'utf', get_string_format(data)

tests/unit/validator_tests/test_validate.py

@@ -4,6 +4,8 @@
 
 import h5py
 import numpy as np
+import pytest
+from hdmf.validate.validator import get_type, EmptyArrayError
 from dateutil.tz import tzlocal
 from hdmf.build import GroupBuilder, DatasetBuilder, LinkBuilder, ReferenceBuilder, TypeMap, BuildManager
 from hdmf.spec import (GroupSpec, AttributeSpec, DatasetSpec, SpecCatalog, SpecNamespace,
@@ -1907,3 +1909,9 @@ def test_isodatetime_no_time_component_fails(self):
                 # This confirms it fails because it lacks the 'T' and timezone
                 self.assertEqual(len(result), 1)
                 self.assertIsInstance(result[0], Error)
+
+def test_get_type_empty_data():
+    """Test that get_type handles empty data with an informative error (Issue #775)."""
+    assert get_type(None) == (None, None)
+    with pytest.raises(EmptyArrayError, match="Dataset is empty; cannot determine type."):
+        get_type([])