Python，Pandas数据框：为什么我不能直接将数据分配给df.values = df.values / 100？

Question

I want to 1) Read in data from the FamaFrench website 2) Convert the date (Month,Year) into a DateTime object 3) Convert all the returns data into percentage returns (returns/100)

My code below reads in data from the famafrench website.

industry30Web =web.DataReader("30_Industry_Portfolios","famafrench")
industry30_monthlyVal = industry30Web[4]
dateInt = industry30_monthlyVal.index
conv = lambda x: datetime.datetime.strptime(str(x),'%Y%m')
dates = [conv(x) for x in dateInt]
industry30_monthlyVal.index =  dates
industry30_monthlyVal.values = industry30_monthlyVal.values/100 

The last line is showing an AttributeError

Please help and let me know what I'm doing wrong.

Answer 1

The documentation specifically states that you cannot assign to the values attribute.

However, you can achieve what you want by doing industry30_monthlyVal[:]= industry30_monthlyVal[:]/100

Answer 2

When i had a look at the source under generic of pd.DataFrame i found:

@property
def values(self):
    """Numpy representation of NDFrame

    Notes
    -----
    The dtype will be a lower-common-denominator dtype (implicit
    upcasting); that is to say if the dtypes (even of numeric types)
    are mixed, the one that accommodates all will be chosen. Use this
    with care if you are not dealing with the blocks.

    e.g. If the dtypes are float16 and float32, dtype will be upcast to
    float32.  If dtypes are int32 and uint8, dtype will be upcase to
    int32.
    """
    return self.as_matrix()

The method has no capability for writing data to the DataFrame (like ie an class variable that you can overwrite). The apply functionality is probably what your are looking for:

import numpy as np, pandas as pd
s = pd.dataFrame(np.random.randn(10))
s = s.apply(lambda x: x/100)

It works the same way for a data frame