Merge pull request #139 from fact-project/h5py3

Fix string conversion for h5py >= 3

fact/VERSION

@@ -1 +1 @@
-0.26.1
+0.26.2

fact/io.py

@@ -177,9 +177,9 @@ def read_h5py(
             dataset = group[col]
             array = to_native_byteorder(dataset[first:last])
 
-            # pandas cannot handle bytes, convert to str
-            if array.dtype.kind == 'S':
-                array = array.astype(str)
+            # decode unicode strings to str
+            if array.dtype.kind in {'S', 'O'}:
+                array = array.astype('U')
 
             if parse_dates and dataset.attrs.get('timeformat') is not None:
                 array = pd.to_datetime(array, infer_datetime_format=True)

tests/test_io.py

@@ -295,7 +295,8 @@ def test_read_data_h5py():
 
     df = pd.DataFrame({
         'x': np.random.normal(size=50).astype('float32'),
-        'N': np.random.randint(0, 10, dtype='uint8', size=50)
+        'N': np.random.randint(0, 10, dtype='uint8', size=50),
+        'name': [f"s{i}" for i in range(50)],
     }).sort_index(1)
 
     with tempfile.NamedTemporaryFile(suffix='.hdf5') as f: