Imputing values from list into Pandas data frame

Question

I have a Pandas data frame that I would like to update a column. Currently the format is like with many many lines. If the value equals DI would like to random choose from a list to replace that value with. For example:

Values
A
B
C
D

my_list = ["E", "F", "G"]

df['Values'] = pd.np.where(df['Values'].str.contains("D"), random.choice(my_list), df['Values'])

When I do this it only grabs one value let's say "F" and replaces all of "D". I would like to go row by row to distribute randomly. So for example if I am replacing 100 D's I might get, 40 "E's" 25 F's and 35 G's. Any thoughts on how I can tweak this?

Thanks!

Answer 1

You can assign

m = df['Values'].str.contains("D")

df.loc[m,'Values']=np.random.choice(my_list,m.sum())
df
Out[27]: 
  Values
0      A
1      B
2      F
3      E