Python Pandas Dataframe: change a NaN cell value with a different column from previous row

Question

import pandas as pd
import numpy as np

data = np.array([['', 'Col1', 'Col2', 'Col3'],
                 ['Row1', 1, 2, 3],
                 ['Row2', np.nan, 5, 6],
                 ['Row3', 7, 8, 9]
                 ])

df = pd.DataFrame(data=data[1:, 1:],
                  index=data[1:,0],
                  columns=data[0,1:])


OutPut:
     Col1 Col2 Col3
Row1    1    2    3
Row2  nan    5    6
Row3    7    8    9

I would like to loop through the dataframe and replace the NaN value in Row2['Col1'] (current row in loop) with the value in Row1['Col3'] (different column from the previous record in loop)

Answer 1

One way you can do this is to use stack , ffill , and unstack :

df.stack(dropna=False).ffill().unstack()

Output:

     Col1 Col2 Col3
Row1    1    2    3
Row2    3    5    6
Row3    7    8    9

Answer 2

You have one more thing need to solve before replace nan :

1st: You are using array , array do not accept join type , which mean your nan here is not np.nan any more, it is 'nan'

df.applymap(type)
Out[1244]: 
               Col1           Col2           Col3
Row1  <class 'str'>  <class 'str'>  <class 'str'>
Row2  <class 'str'>  <class 'str'>  <class 'str'>
Row3  <class 'str'>  <class 'str'>  <class 'str'>

df=df.replace('nan',np.nan)

2nd, I am using np.roll + combine_first to fill the nan

df.combine_first(pd.DataFrame(np.roll(np.concatenate(df.values),1).reshape(3,3),index=df.index,columns=df.columns))
Out[1240]: 
     Col1 Col2 Col3
Row1    1    2    3
Row2    3    5    6
Row3    7    8    9

Answer 3

I apologize for not posting the actual data from my dataset so here it is:

             Open   High    Low   Last  Change  Settle   Volume  
Date                                                              
2017-05-22  51.97  52.28  51.73  **51.96**    0.49   52.05  70581.0   
2017-05-23    **NaN**  52.44  51.61  52.31    0.24   52.35   9003.0   
2017-05-24  52.34  52.63  51.91  52.05    0.23   52.12  11678.0   
2017-05-25  52.25  52.61  49.49  49.59    2.28   49.84  19721.0   
2017-05-26  49.82  50.73  49.34  50.73    0.82   50.66  11214.0 

I needed the script to find any 'NaN's in the 'Open' column and replace it with the 'Last' from the previous row.(highlighted here by double asterisks).

I thank all for the posts, however, this is what ended up working:

missing = df['Open'].isnull() # get nans
new_open = df['Open'].copy() # make copy

# loop missing and test against a True value
# if so, get the 'Last' value at index and
# populate new_open value at index
for i in range(missing.shape[0]):
    if missing[i] == True:
        new_open.iloc[i] = df['Last'][i-1]

# replace the 'Open' values with new 'Open' values
df['Open'] = new_open