Merge pull request #11 from schrimpf/fix_amin

fix for pandas.aggregate changing name of created variable
bernardodionisi · Dec 9, 2023 · 9e429e5 · 9e429e5
2 parents 63b4bc2 + 3013a73
commit 9e429e5
Showing 1 changed file with 8 additions and 5 deletions.
diff --git a/differences/tools/panel_validation.py b/differences/tools/panel_validation.py
@@ -133,9 +133,12 @@ def __set__(self, instance, data):
             )
 
             # for each entity-event_date,
-            # create the start (amin) and end (amax) of the panel for the corresponding entity
+            # create the start (_min_time) and end (_max_time) of the panel for the corresponding entity
             cohort_data = cohort_data[cohort_info_list].join(
-                data.groupby(entity_name)[time_name].agg([np.min, np.max])
+                data
+                .groupby(entity_name)[time_name]
+                .agg(['min', 'max'])
+                .set_axis(labels=['_min_time', '_max_time'], axis=1)
             )
 
             # ----------- pre-process cohorts / data -------------------
@@ -330,7 +333,7 @@ def pre_process_treated_before(
 
     # entities whose event happened BEFORE the start of their time
     treated_before = cohort_data.loc[
-        lambda x: x[cohort_name] <= x["amin"]
+        lambda x: x[cohort_name] <= x["_min_time"]
     ].index.unique()
 
     if len(treated_before):
@@ -357,12 +360,12 @@ def pre_process_treated_after(
     """drops event dates that come after the end of the panel (for the entity)"""
 
     # entities whose event happened AFTER the end of their time
-    bool_after_end = cohort_data[cohort_name] > cohort_data["amax"]
+    bool_after_end = cohort_data[cohort_name] > cohort_data["_max_time"]
 
     # shouldn't this be [cohort - anticipation] ?
     # line 85: https://github.com/bcallaway11/did/blob/master/R/pre_process_did.R
     # bool_after_end = (cohort_data[cohort_name] -
-    #                   anticipation > cohort_data['amax'])
+    #                   anticipation > cohort_data['_max_time'])
 
     n_entities_after_end = np.sum(bool_after_end)  # number of entities
     if n_entities_after_end: