rolling_apply and dataframes

[denadai2]

Hello, I'm trying to apply this

from scipy.spatial import distance
def myrolling_apply(df):
    print type(df)
    print df
    print df.dtypes
    a = df['predicted'].values
    b = df['actual'].values
    #dst = distance.euclidean(a,b)
    return(df)

rrbug = pd.read_csv('bug.csv')
rrbug = pd.rolling_apply(rrbug, 24, myrolling_apply)

To this dataframe https://dl.dropboxusercontent.com/u/1118905/bug.csv. I would aspect to have a dataframe BUT df is a ndarray, as you see from the results:

<type 'numpy.ndarray'>
[ 18.038866  18.81628   15.652594  19.418134  23.685213  21.636703
  32.264257  44.563431  63.160193  46.759563  44.06377   45.75921
  57.929419  58.121308  61.65729   59.300808  64.033246  57.590802
  56.642796  53.321256  68.65092   57.555255  39.222144  34.622948]

[jreback]

this is a dupe of #5071

currently a 2-d ndarray is passed. Pls feel free to submit a pull-request to enhance. (You can easily just wrap this in a DataFrame if you want inside your function).