Unable to make float from 'nan' pandas dataframe

Asked Jul 29 '17 at 10:48

Active Jul 29 '17 at 10:48

Viewed 269 times

I am very much new in Pandas . My code snippet is below

filtered_sales_user = df_sales_User.drop('custid',axis=1).groupby('runseq').mean()

Basically I am grouping dataframe by runseq and calculating the mean.

But for a dataframe below, where few columns are NaN I am getting below error.

custid        m1    m2  m3  m4  m5       m12       m13       m14  \
0    4894  0.637958  0.01 NaN NaN NaN  0.189322  0.211065  0.002590
1    8080  0.854400   NaN NaN NaN NaN  0.009367  0.065700  0.012541

        m6      m7  b1  b3           b40   b41  runseq  net_sales
0  Money_Funds  Growth NaN NaN  3.171390e+06  18.0      29        NaN
1  Money_Funds    None NaN NaN  5.551656e+05  15.0      29        NaN

Error from server: code=2200 [Invalid query] message="Unable to make float from 'nan'"

Now should I replace all the NaN using fillna to fix this issue. And also will it hamper my mean calculation .

asked Jul 29 '17 at 10:48

curiousguy

3,212
8
39
71

You can use `dropna` instead, so your mean wouldn't be affected – taras Jul 29 '17 at 10:56
So will it still calculate mean for the above example , where say for `m3` all values are `NaN`. What will be mean for `m3` then – curiousguy Jul 29 '17 at 10:59
https://stackoverflow.com/questions/40299055/pandas-how-to-fill-null-values-with-mean-of-a-groupby – taras Jul 29 '17 at 11:00
Replace nan values with average of columns is not working . Because some columns consists NaN only . – curiousguy Jul 29 '17 at 11:14
Just drop those columns. – taras Jul 29 '17 at 11:15
No I cant drop actually I am storing them for reference. So drop means no records and might cause anomaly . – curiousguy Jul 29 '17 at 11:17

Unable to make float from 'nan' pandas dataframe

0 Answers0