阅读中文版本

pandas SettingWithCopyWarning / Chained Assignment

Eliminate pandas SettingWithCopyWarning and FutureWarning on chained assignment by using .loc on a real DataFrame copy instead of slicing.

python -c "import pandas as pd; df=pd.DataFrame({'a':[1,-1]}); df[df.a>0]['b']=2; print(df)"

df = df.copy(); df.loc[df.a > 0, 'b'] = 2

Symptoms

When you hit “pandas SettingWithCopyWarning / Chained Assignment”, guessing wastes time. The guide below follows evidence: reproduce the symptom, run targeted diagnostics, identify the root cause, apply the fix, and verify it with concrete checks.

Observed symptoms include:

• pandas logs “SettingWithCopyWarning: A value is trying to be set on a copy of a slice”
• A filtered column assignment silently does not persist
• pandas 2.2 shows “FutureWarning: ChainedAssignmentError”

Reproduce and Diagnose

Reproduce the failure in isolation before changing any code. Work through the checks below in order; each one narrows the search space.

Step 1: Reproduce with a single-chain assignment

python -c "import pandas as pd; df=pd.DataFrame({'a':[1,-1]}); df[df.a>0]['b']=2; print(df)"

The warning appears and column b is missing, confirming the write was lost.

Step 2: Rewrite with .loc on a copy

df = df.copy(); df.loc[df.a > 0, 'b'] = 2

A single .loc call on an owned DataFrame assigns reliably and silences the warning.

Likely contributors, in order of frequency:

• Assignment targets a slice such as df[df.a > 0][“b”] = 1
• The DataFrame is a view created by filtering without an explicit copy
• Enlargement through chained indexing may write to a temporary

Root Cause

Chained indexing performs two operations: a getitem that may return a view or a copy, then a setitem on that result. When the intermediate is a copy, the assignment mutates the temporary and is lost; pandas warns because it cannot guarantee intent.

The Fix

Use a single .loc[row_indexer, col] = value on an explicitly owned DataFrame (df = df.copy() when derived from a filter), and read with .loc or assignment via df.assign to avoid chains.

Watch out: Adding a copy can raise memory use on large frames; when performance matters, select columns you need instead of copying everything.

import pandas as pd

df = pd.DataFrame({'a': [1, -1], 'b': [0, 0]})

# ❌ chained assignment (warning, write may be lost)
# df[df.a > 0]['b'] = 2

# ✅ single .loc on an owned frame
filtered = df.copy()
filtered.loc[filtered.a > 0, 'b'] = 2
print(filtered)

Verify and Prevent

Rerun the assignment and confirm the new column exists and the warning is gone; run with -W error::FutureWarning to make pandas future errors fatal in CI.

Once verified, add a regression test that reproduces the original symptom so the failure cannot silently return.

This class of failure is usually environmental rather than pure logic, so capture the exact command, the environment variables, and the working directory in your bug report. Reproduce it in the smallest possible script, then widen the scope only once the minimal case is understood. A short written checklist for this failure mode saves hours the next time it appears.