Pandas itertuples返回结果为何不是预期的named tuple？

This is totally normal behavior, not an error on your part! Here's what's going on:

• When you call df.itertuples(), pandas returns objects that act exactly like named tuples, but they might be instances of pandas.core.frame.Synonym instead of Python's standard namedtuple—depending on your CSV column names.
• Synonym is actually a subclass of Python's built-in namedtuple, so it has all the same functionality: you can access fields by name (with spaces converted to underscores), unpack the tuple, or access elements by index.
• The switch to Synonym happens when your column names have characters that aren't valid Python identifiers—things like spaces, hyphens, or names starting with a number. Even with just two columns, if their names fall into this category, pandas uses Synonym to handle them while keeping the named tuple interface intact.

Example Check #

You can confirm that Synonym behaves like a named tuple with this quick test:

from collections import namedtuple
row = next(df.itertuples())
print(isinstance(row, namedtuple))  # Will output True

How to Get Standard Named Tuples (If You Want) #

If you specifically need regular namedtuple instances, just rename your columns to valid Python identifiers first:

# Rename columns to meaningful valid names (e.g., replace spaces with underscores)
df.columns = ['column_one', 'column_two']
# Convert to list of named tuples (exclude index with index=False if needed)
named_tuple_list = list(df.itertuples(index=False))

Either way, whether you're dealing with Synonym or a standard named tuple, you can work with the rows the same way—so no need to worry about breaking your code!

内容的提问来源于stack exchange，提问作者max