每个单独的行和列（每个字段）随机播放逗号分隔的字符串

Question

I have a dataframe with rows of values that have been concatenated, but separated by a comma.

Row1 foo,bar,test,case

Row2 base,ball,basket,foot

The goal is the shuffle/randomize each fields values, will retaining row order (do not shuffle columns, index must be kept) The hope is to return something like this:

Row1 test,foo,case,bar

Row2 ball,foot,base,basket

SOLUTION FOUND:

Original_DF = # Our csv loaded data - the DF contains multiple columns of data attached to primary

data_list=[e for e in Original_DF['Data_List']]   # each 'Data_List' field was one long string with a comma seperating words, we needed to make them a list

Shuff_DF=pd.DataFrame()
for i in range(len(data_list)):
    myList=np.random.permutation(data_list[i].split(","))
    myString = ",".join(myList)
    Shuff_DF = Shuff_DF.append({'Data_List2': myString}, ignore_index=True)

Original_DF['Data_List2']=Shuff_DF['Data_List2'] # Append newly shuffled Lists to original df

Answer 1

You can use numpy.random.permutation to shuffle a list

https://docs.scipy.org/doc/numpy-1.14.0/reference/generated/numpy.random.permutation.html

Example:

import numpy.random

mydata = "foo,bar,baz,bat"

print(numpy.random.permutation(mydata.split(",")))

Answer 2

Another approach, using pandas functionality (sample):

df = pd.DataFrame({"a": ["foo", "bar", "test", "case"], 
                   "b": ["xoo", "xar", "xest", "xase"]})
for col in df.columns:
    df[col] = df[col].sample(frac=1).values

Probably there is a more elegant approach avoiding the for loop, using apply.

PS: Alternatively, modifying the other answer from @Simon Crane:

df.apply(np.random.permutation, axis=0)