From 40a93ef813767d633ebfa350a15d73c636326dd9 Mon Sep 17 00:00:00 2001
From: John Bogaardt <jbogaardt@wcf.com>
Date: Thu, 13 Feb 2025 08:29:06 -0700
Subject: [PATCH] 0.8.24 bump

---
 chainladder/core/base.py     |  4 ++--
 chainladder/core/dunders.py  | 15 +++++----------
 chainladder/core/triangle.py |  7 +++++++
 requirements.txt             |  4 ++--
 4 files changed, 16 insertions(+), 14 deletions(-)

diff --git a/chainladder/core/base.py b/chainladder/core/base.py
index f3e9401a..122e8784 100644
--- a/chainladder/core/base.py
+++ b/chainladder/core/base.py
@@ -90,14 +90,14 @@ def _aggregate_data(data, origin_date, development_date, index, columns):
             key_gr = ["__origin__", "__development__"] + [
                 data[item] for item in ([] if not index else index)
             ]
-            data_agg = data.groupby(key_gr)[columns].sum().reset_index().fillna(0)
+            data_agg = data.groupby(key_gr)[columns].sum(numeric_only=False).reset_index().fillna(0)
             data = data.drop(["__origin__", "__development__"], axis=1)
         else:
             # Summarize dataframe to the level specified in axes
             key_gr = [origin_date, development_date] + [
                 data[item] for item in ([] if not index else index)
             ]
-            data_agg = data[columns].groupby(key_gr).sum().reset_index().fillna(0)
+            data_agg = data[columns].groupby(key_gr)[columns].sum(numeric_only=False).reset_index().fillna(0)
             data_agg["__origin__"] = data_agg[origin_date.name]
             data_agg["__development__"] = data_agg[development_date.name]
         # origin <= development is required - truncate bad records if not true
diff --git a/chainladder/core/dunders.py b/chainladder/core/dunders.py
index 0b987985..137f1dd7 100644
--- a/chainladder/core/dunders.py
+++ b/chainladder/core/dunders.py
@@ -109,14 +109,10 @@ def _prep_columns(self, x, y):
             y.vdims = x.vdims
         elif x.shape[1] == y.shape[1] and np.array_equal(x.columns, y.columns):
             return x, y
-        else:
-            # Use sets for faster operations
-            x_cols = set(x.columns)
-            y_cols = set(y.columns)
-            
+        else:            
             # Find columns to add to each triangle
-            cols_to_add_to_x = y_cols - x_cols
-            cols_to_add_to_y = x_cols - y_cols
+            cols_to_add_to_x = [col for col in y.columns if col not in x.columns] 
+            cols_to_add_to_y = [col for col in x.columns if col not in y.columns] 
             
             # Create new columns only if necessary
             if cols_to_add_to_x:
@@ -128,9 +124,8 @@ def _prep_columns(self, x, y):
                 y = y.reindex(columns=new_y_cols, fill_value=0)
             
             # Ensure both triangles have the same column order
-            final_cols = list(x_cols | y_cols)
-            x = x[final_cols]
-            y = y[final_cols]
+            x = x[new_x_cols]
+            y = y[new_x_cols]
         
         # Reset backends only if they've changed
         if x.array_backend != x_backend:
diff --git a/chainladder/core/triangle.py b/chainladder/core/triangle.py
index 0458b774..55303c32 100644
--- a/chainladder/core/triangle.py
+++ b/chainladder/core/triangle.py
@@ -981,3 +981,10 @@ def sort_axis(self, axis):
                 obj.values = obj.values[..., list(sort)]
                 obj.ddims = obj.ddims[list(sort)]
         return obj
+
+    def reindex(self, columns=None, fill_value=np.nan):
+        obj = self.copy()
+        for column in columns:
+            if column not in obj.columns:
+                obj[column] = fill_value
+        return obj
diff --git a/requirements.txt b/requirements.txt
index e2d6a6fc..b6f6cb7c 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,5 +1,5 @@
-pandas>=0.23
-scikit-learn>=0.23
+pandas>=2.0
+scikit-learn>1.4.2
 numba>0.54
 sparse>=0.9
 matplotlib