[Python] Alphalens get_clean_factor_and_forward_returns throws Length mismatch: Expected axis...

I'm using alphalens.utils.get_clean_factor_and_forward_returns() to compute forward returns from a factor series and price DataFrame. But I'm hitting the following error:

ValueError: Length mismatch: Expected axis has 34 elements, new values have 36 elements


My Input Data:

• 
  factor_series: A Pandas Series with a MultiIndex of (date, asset), total of 37 rows Example:
  
  date asset 2007-01-01 AAPL 0.503 AMZN 0.941 2008-01-01 AAPL 0.900 AMZN 0.993 ... 2025-01-01 AMZN 1.000 Name: factor, dtype: float64
  
  
• 
  prices: A DataFrame with DatetimeIndex and tickers as columns.
  Shape: (64 dates, 2 tickers — AAPL and AMZN)
  Note: AAPL is missing data on the last date 2025-01-01.
  

Here’s a minimal version of my code that reproduces the issue:

import pandas as pd
from alphalens.utils import get_clean_factor_and_forward_returns

# Load factor data (CSV: date, asset, factor)
factor_df = pd.read_csv("factor_series_Item1A_positive.csv")
prices_df = pd.read_csv("prices_filtered_Item1A_positive.csv", index_col=0)

# Convert to proper datetime format
factor_df['date'] = pd.to_datetime(factor_df['date'])
prices_df.index = pd.to_datetime(prices_df.index)

# Convert factor DataFrame to MultiIndex Series
factor_series = factor_df.set_index(['date', 'asset'])['factor']
factor_series.name = 'factor'

# Align assets
common_assets = factor_series.index.get_level_values("asset").unique().intersection(prices_df.columns)
factor_series = factor_series[factor_series.index.get_level_values("asset").isin(common_assets)]
prices_df = prices_df[common_assets]

# Align dates
common_dates = factor_series.index.get_level_values("date").unique().intersection(prices_df.index)
factor_series = factor_series[factor_series.index.get_level_values("date").isin(common_dates)]
prices_df = prices_df.loc[common_dates]

# Remove dates that can't have forward returns (e.g., last date)
last_valid_date = prices_df.index.max()
factor_series = factor_series[factor_series.index.get_level_values("date") < last_valid_date]

# Attempt to compute forward returns
data = get_clean_factor_and_forward_returns(
factor_series,
prices_df,
quantiles=5,
bins=None,
periods=[1]
)


What I've Tried:

• 
  Verified that all dates and tickers in factor_series are in prices
  
  
• 
  Filtered factor_series to exclude the last date (where forward returns aren't possible)
  
  
• 
  Ensured both inputs are properly sorted and have no NaNs
  
  
• 
  Even after filtering to 36 rows in factor_series, I still get this error
  

Question:

Where exactly in get_clean_factor_and_forward_returns() does this length mismatch occur?
What specifically is mismatching — quantile bin assignment, forward return calculation, or an internal reindex?
How can I fix the issue?

Continue reading...