Dead code in TriangleDunders._prep_columns()

Are you on the latest chainladder version?

Yes, this bug occurs on the latest version.

Describe the bug in words

Lines 147 - 150 can't be reached:

chainladder-python/chainladder/core/dunders.py

Lines 114 to 152 in 6f80200

    
           def _prep_columns(self, x, y): 
        
               x_backend, y_backend = x.array_backend, y.array_backend 
        
               if len(x.columns) == 1 and len(y.columns) > 1: 
        
                   x.vdims = y.vdims 
        
               elif len(y.columns) == 1 and len(x.columns) > 1: 
        
                   y.vdims = x.vdims 
        
               elif len(y.columns) == len(x.columns) == 1 and x.columns != y.columns: 
        
                   y.vdims = x.vdims 
        
               elif x.shape[1] == y.shape[1] and np.array_equal(x.columns, y.columns): 
        
                   return x, y 
        
               else:             
        
                   # Find columns to add to each triangle 
        
                   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] 
        
                   # Start with case with no new columns, y simply has a different order 
        
                   new_x_cols = list(x.columns) 
        
                   # Then, if there are new columns, add them. 
        
                   if cols_to_add_to_x: 
        
                       new_x_cols = list(x.columns) + list(cols_to_add_to_x) 
        
                       x = x.reindex(columns=new_x_cols, fill_value=0) 
        
                   if cols_to_add_to_y: 
        
                       new_y_cols = list(y.columns) + list(cols_to_add_to_y) 
        
                       y = y.reindex(columns=new_y_cols, fill_value=0) 
        
                   # Ensure both triangles have the same column order 
        
                   x = x[new_x_cols] 
        
                   y = y[new_x_cols] 
        
               # Reset backends only if they've changed 
        
               if x.array_backend != x_backend: 
        
                   x = x.set_backend(x_backend, inplace=True) 
        
               if y.array_backend != y_backend: 
        
                   y = y.set_backend(y_backend, inplace=True) 
        
               return x, y

How can the bug be reproduced?

This was discovered when trying to write unit tests for them. This diagnostic was generated by Claude, so I want to double check whether it's correct before we remove the lines:

● These lines are dead code — the backend never changes inside _prep_columns. Here's why:

  1. _prep_columns is always called after _compatibility_check, which calls set_common_backend to align both triangles to the same backend before we get here.
  2. None of the operations inside _prep_columns (vdims assignment, reindex, column slicing) change the backend — I verified this empirically with a monkey-patch that would detect any change.
  So x.array_backend != x_backend and y.array_backend != y_backend are always False, and lines 146/148 can never execute. There's no test that can reach them.

What is the expected behavior?

My IDE detects that there's only 1 usage of TriangleDunders._prep_columns(), and it does occur after _compatibility_check(), so item 1 is correct. We just need to check item 2 before we remove the lines.

Would you be willing to contribute this ticket?

Yes, absolutely!
Yes, but I would like some help.
No.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Dead code in TriangleDunders._prep_columns() #1045

Are you on the latest chainladder version?

Describe the bug in words

How can the bug be reproduced?

What is the expected behavior?

Would you be willing to contribute this ticket?

Metadata

Assignees

Labels

Type

Fields

Projects

Milestone

Relationships

Development

	def _prep_columns(self, x, y):
	x_backend, y_backend = x.array_backend, y.array_backend

	if len(x.columns) == 1 and len(y.columns) > 1:
	x.vdims = y.vdims
	elif len(y.columns) == 1 and len(x.columns) > 1:
	y.vdims = x.vdims
	elif len(y.columns) == len(x.columns) == 1 and x.columns != y.columns:
	y.vdims = x.vdims
	elif x.shape[1] == y.shape[1] and np.array_equal(x.columns, y.columns):
	return x, y
	else:
	# Find columns to add to each triangle
	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]

	# Start with case with no new columns, y simply has a different order
	new_x_cols = list(x.columns)

	# Then, if there are new columns, add them.
	if cols_to_add_to_x:
	new_x_cols = list(x.columns) + list(cols_to_add_to_x)
	x = x.reindex(columns=new_x_cols, fill_value=0)

	if cols_to_add_to_y:
	new_y_cols = list(y.columns) + list(cols_to_add_to_y)
	y = y.reindex(columns=new_y_cols, fill_value=0)

	# Ensure both triangles have the same column order
	x = x[new_x_cols]
	y = y[new_x_cols]

	# Reset backends only if they've changed
	if x.array_backend != x_backend:
	x = x.set_backend(x_backend, inplace=True)
	if y.array_backend != y_backend:
	y = y.set_backend(y_backend, inplace=True)

	return x, y

Search code, repositories, users, issues, pull requests...

Uh oh!

Dead code in TriangleDunders._prep_columns() #1045

Description

Are you on the latest chainladder version?

Describe the bug in words

How can the bug be reproduced?

What is the expected behavior?

Would you be willing to contribute this ticket?

Metadata

Metadata

Assignees

Labels

Type

Fields

Projects

Milestone

Relationships

Development

Issue actions