Outputs for README example don't match

[jseabold]

There doesn't seem to be a __version__ in the code, but I installed via pip semi-recently. The filtered_output and the pandas-ply output in the README don't match. The pandas-ply results are missing January. On Python 3.4

import pandas as pd
from ply import install_ply, X
install_ply(pd)

%load_ext rpy2.ipython.rmagic
from pandas.rpy import common as com
%R library("nycflights13")
flights = com.load_data("flights")

grouped_flights = flights.groupby(['year', 'month', 'day'])
output = pd.DataFrame()
output['arr'] = grouped_flights.arr_delay.mean()
output['dep'] = grouped_flights.arr_delay.mean()
filtered_output = output[(output.arr > 30) & (output.dep > 30)]

print(filtered_output)

(flights
  .groupby(['year', 'month', 'day'])
  .ply_select(
    arr = X.arr_delay.mean(),
    dep = X.dep_delay.mean())
  .ply_where(X.arr > 30, X.dep > 30))

Produces

[42]: print(filtered_output)
                      arr        dep
year month day                      
2013 1     16   34.247362  34.247362
           31   32.602854  32.602854
     2     11   36.290094  36.290094
           27   31.252492  31.252492
     3     8    85.862155  85.862155
           18   41.291892  41.291892
     4     10   38.412311  38.412311
           12   36.048140  36.048140
           18   36.028481  36.028481
           19   47.911697  47.911697
           22   37.812166  37.812166
           25   33.681250  33.681250
     5     8    39.609183  39.609183
           23   61.970899  61.970899
     6     13   63.753689  63.753689
           18   37.648026  37.648026
           24   51.176808  51.176808
           25   41.513684  41.513684
           27   44.783296  44.783296
           28   44.976852  44.976852
           30   43.510278  43.510278
     7     1    58.280502  58.280502
           7    40.306378  40.306378
           9    31.334365  31.334365
           10   59.626478  59.626478
           22   62.763403  62.763403
           23   44.959821  44.959821
           28   49.831776  49.831776
     8     1    35.989259  35.989259
           8    55.481163  55.481163
           9    43.313641  43.313641
           28   35.203074  35.203074
     9     2    45.518430  45.518430
           12   58.912418  58.912418
     10    7    39.017260  39.017260
     12    5    51.666255  51.666255
           8    36.911801  36.911801
           9    42.575556  42.575556
           10   44.508796  44.508796
           14   46.397504  46.397504
           17   55.871856  55.871856
           23   32.226042  32.226042

and

                      dep        arr
year month day                      
2013 2     11   39.073598  36.290094
           27   37.763274  31.252492
     3     8    83.536921  85.862155
           18   30.117960  41.291892
     4     10   33.023675  38.412311
           12   34.838428  36.048140
           18   34.915361  36.028481
           19   46.127828  47.911697
           22   30.642553  37.812166
     5     8    43.217778  39.609183
           23   51.144720  61.970899
     6     13   45.790828  63.753689
           18   35.950766  37.648026
           24   47.157418  51.176808
           25   43.063025  41.513684
           27   40.891232  44.783296
           28   48.827784  44.976852
           30   44.188179  43.510278
     7     1    56.233825  58.280502
           7    36.617450  40.306378
           9    30.711499  31.334365
           10   52.860702  59.626478
           22   46.667047  62.763403
           23   44.741685  44.959821
           28   37.710162  49.831776
     8     1    34.574034  35.989259
           8    43.349947  55.481163
           9    34.691898  43.313641
           28   40.526894  35.203074
     9     2    53.029551  45.518430
           12   49.958750  58.912418
     10    7    39.146710  39.017260
     12    5    52.327990  51.666255
           9    34.800221  42.575556
           17   40.705602  55.871856
           23   32.254149  32.226042

[jhorowitz-coursera]

@jseabold -- Agh, great catch. There's a typo in the readme. It says:

output['arr'] = grouped_flights.arr_delay.mean()
output['dep'] = grouped_flights.arr_delay.mean()

but it should say

output['arr'] = grouped_flights.arr_delay.mean()
output['dep'] = grouped_flights.dep_delay.mean()

With that fix, the with- and without-ply fragments give the same output (up to column order, which the keyword-argument form of ply_select doesn't control). I'll fix that up. Thanks!

[jhorowitz-coursera]

Fixed in readme & docs; I'll push the change to the docs when I release a new version. (I'll also add a __version__, as you mentioned would be useful.)