pd.DataFrame.expanding bug with axis = 1

[boonteck]

import pandas as pd
import numpy as np

df = pd.DataFrame(np.ones((10, 20)))
df.expanding(3, axis = 1).sum()
   0   1    2    3    4    5    6    7    8     9     10    11    12    13  \
0 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
1 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
2 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
3 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
4 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
5 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
6 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
7 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
8 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   
9 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  10.0  10.0  10.0  10.0   

     14    15    16    17    18    19  
0  10.0  10.0  10.0  10.0  10.0  10.0  
1  10.0  10.0  10.0  10.0  10.0  10.0  
2  10.0  10.0  10.0  10.0  10.0  10.0  
3  10.0  10.0  10.0  10.0  10.0  10.0  
4  10.0  10.0  10.0  10.0  10.0  10.0  
5  10.0  10.0  10.0  10.0  10.0  10.0  
6  10.0  10.0  10.0  10.0  10.0  10.0  
7  10.0  10.0  10.0  10.0  10.0  10.0  
8  10.0  10.0  10.0  10.0  10.0  10.0  
9  10.0  10.0  10.0  10.0  10.0  10.0  

axis = 1 on pd.DataFrame.expanding doesn't seem to be working as expected; a horizontal expanding window. Instead, the window size appears to be constrained to the number of rows.

axis = 0 is working fine as a vertical expanding window

Expected Output

df.T.expanding(3, axis = 0).sum().T

  0   1    2    3    4    5    6    7    8     9     10    11    12    13  \
0 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
1 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
2 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
3 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
4 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
5 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
6 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
7 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
8 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   
9 NaN NaN  3.0  4.0  5.0  6.0  7.0  8.0  9.0  10.0  11.0  12.0  13.0  14.0   

     14    15    16    17    18    19  
0  15.0  16.0  17.0  18.0  19.0  20.0  
1  15.0  16.0  17.0  18.0  19.0  20.0  
2  15.0  16.0  17.0  18.0  19.0  20.0  
3  15.0  16.0  17.0  18.0  19.0  20.0  
4  15.0  16.0  17.0  18.0  19.0  20.0  
5  15.0  16.0  17.0  18.0  19.0  20.0  
6  15.0  16.0  17.0  18.0  19.0  20.0  
7  15.0  16.0  17.0  18.0  19.0  20.0  
8  15.0  16.0  17.0  18.0  19.0  20.0  
9  15.0  16.0  17.0  18.0  19.0  20.0  

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.7.1.final.0 python-bits: 64 OS: Linux OS-release: 4.17.19-1-MANJARO machine: x86_64 processor: byteorder: little LC_ALL: None LANG: en_SG.UTF-8 LOCALE: en_SG.UTF-8

pandas: 0.23.4
pytest: None
pip: 18.0
setuptools: 40.4.3
Cython: 0.29
numpy: 1.15.2
scipy: 1.1.0
pyarrow: None
xarray: None
IPython: 6.5.0
sphinx: None
patsy: 0.5.0
dateutil: 2.7.3
pytz: 2018.5
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.2.3
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: 1.0.1
sqlalchemy: 1.2.12
pymysql: None
psycopg2: 2.7.5 (dt dec pq3 ext lo64)
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[gfyoung]

That looks very odd indeed! Investigation and PR are welcome!

[jreback]

its using the wrong axis for computing the window length, IOW its always using axis=0

see Expanding._get_window, should be easy patch.