Why Is the replace() Method Not Replacing Values in My Pandas DataFrame?

Despite attempting to replace specific strings with NaN in a simple DataFrame, the replace() method appears to be ineffective. For instance:

d = {'color' : pd.Series(['white', 'blue', 'orange']),
   'second_color': pd.Series(['white', 'black', 'blue']),
   'value' : pd.Series([1., 2., 3.])}
df = pd.DataFrame(d)
df.replace('white', np.nan)

The expected output, with 'white' values replaced by NaN, is not achieved. Instead, the DataFrame remains unchanged.

A Solution: Enable Partial Replacements with regex=True

By default, the replace() method performs full string replacements. To enable partial replacements, the regex parameter must be set to True. This small tweak allows us to replace specific strings anywhere within the DataFrame:

df.replace('white', np.nan, regex=True)

With this modification, the DataFrame will correctly replace all instances of 'white' with NaN.