fixed: value error in select and drop_corr method

CHANGELOG.md

@@ -11,6 +11,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Added `performance` in `toad.utils` for test code performance
 - Added `pickletracer` in `toad.utils` for infer requirements in pickle object
 
+### Fixed
+- Fixed `Value Error` in `select` and `drop_corr` method when using `pandas >= 2.0.x`
+
 ## [0.1.2] - 2023-04-09
 
 ### Add

toad/__init__.py

@@ -10,6 +10,7 @@
 from .detector import detect
 from .metrics import KS, KS_bucket, F1
 from .stats import quality, IV, VIF, WOE, entropy, entropy_cond, gini, gini_cond
+from .transform import Combiner, WOETransformer
 from .selection import select
 from .scorecard import ScoreCard
 from .utils import Progress, performance

toad/selection.py

@@ -326,7 +326,7 @@ def drop_corr(frame, target = None, threshold = 0.7, by = 'IV',
 
     f, t = split_target(frame[cols], target)
 
-    corr = f.corr().abs()
+    corr = f.corr(numeric_only = True).abs()
 
     drops = []
 

toad/selection_test.py

@@ -48,7 +48,14 @@ def test_drop_var_exclude():
 
 def test_drop_corr():
     df = drop_corr(frame, target = 'target')
-    assert ['D', 'E', 'F', 'target'] == df.columns.tolist()
+    assert set(['D', 'E', 'F', 'target']) == set(df.columns.tolist())
+
+def test_drop_corr_with_string():
+    ab = np.array(list('ABCDEFG'))
+    str_feat = pd.Series(ab[np.random.choice(7, 500)])
+
+    df = drop_corr(pd.concat((frame, str_feat.rename('str_f')), axis = 1), target = 'target')
+    assert set(['D', 'E', 'F', 'target', 'str_f']) == set(df.columns.tolist())
 
 def test_drop_iv():
     df = drop_iv(frame, target = 'target', threshold = 0.25)