FIX: fixed dtype of year-only labels

should be int as with versions <0.32.2, not object

Author: gdementen

larray_eurostat/tests/test_tsv.py

@@ -12,3 +12,9 @@ def test_eurostat_get_bad_dataset():
     msg = "HTTP Error 404: Not Found"
     with pytest.raises(HTTPError, match=f'^{re.escape(msg)}$'):
         eurostat_get(dataset)
+
+
+# https://github.com/larray-project/larray_eurostat/issues/33
+def test_year_only_axes_as_int_and_not_object():
+    data = eurostat_get('avia_ec_enterp')
+    assert data.time.dtype.kind == 'i'

larray_eurostat/tsv.py

@@ -1,13 +1,15 @@
 import gzip
 from io import StringIO
 from urllib.request import urlopen
+import numpy as np
 
 from larray import Session, read_eurostat
 
 
 def _remove_chars(s, chars):
     return s.translate({ord(c): None for c in chars})
 
+
 def transform_time_labels(label):
     # Account for ambiguous label type (in multifreq larrays, year is str; in only-year larrays, year is given as int)
     # Need string format to search for patterns (e.g. Q1,M12), but return int format when only years are given.
@@ -24,8 +26,10 @@ def transform_time_labels(label):
     else:
         return str_label
 
+
 EUROSTAT_BASEURL = "https://ec.europa.eu/eurostat/api/dissemination/sdmx/2.1/data/"
 
+
 def _get_one(indicator, *, drop_markers=True):
     """Get one Eurostat indicator and return it as an array."""
     url = f"{EUROSTAT_BASEURL}{indicator}?format=TSV&compressed=true"
@@ -41,7 +45,8 @@ def _get_one(indicator, *, drop_markers=True):
 
             # Rename time axis. Rename time labels and reverse them (compatibility old API)
             la_data = la_data.rename(TIME_PERIOD='time')
-            la_data = la_data.set_labels('time', transform_time_labels)
+            if np.issubdtype(la_data.time.dtype, np.character):
+                la_data = la_data.set_labels('time', transform_time_labels)
             la_data = la_data.reverse('time')
 
             # If only one frequency: subset and return without redundant freq Axis (compatibility old API)