Yield the same order regardless of order in the file and database

sodadata · Nov 28, 2024 · f8c2ac9 · f8c2ac9
1 parent b605610
commit f8c2ac9
Showing 1 changed file with 4 additions and 0 deletions.
diff --git a/soda/scientific/soda/scientific/distribution/comparison.py b/soda/scientific/soda/scientific/distribution/comparison.py
@@ -277,6 +277,10 @@ def evaluate(self) -> dict[str, float]:
         ref_data_frequencies = ref_data_frequencies + 1
         test_data_frequencies = test_data_frequencies + 1
 
+        # sort the data to make sure that the categories are in the same order before cast to array_like
+        ref_data_frequencies = ref_data_frequencies.sort_index()
+        test_data_frequencies = test_data_frequencies.sort_index()
+
         # Normalise because scipy wants sums of observed and reference counts to be equal
         # workaround found and discussed in: https://github.com/UDST/synthpop/issues/75#issuecomment-907137304
         stat_value, p_value = chisquare(