CODEWITHSUNDEEP
pythonpython-3.xpandasnumpydataframe
Ramsey
Ramsey

Reputation: 181

How to replace blank values with previous numbers in pandas

SS     RR
       10.4
12.6  
 
       5.6
8.7

I want to fill in the blank rows with preceding values.

Intended result:

SS     RR
       10.4
12.6   10.4
12.6   10.4      
12.6   5.6
8.7    5.6
8.7    5.6

I am trying to first replace the blank values with NaN and then the iteration of preceding values:

df[df['SS']==""] = np.NaN
df[df['RR']==""] = np.NaN

df.SS.fillna(method='ffill')
df.RR.fillna(method='ffill')

But for some reason it doesn't make any difference to the dataframe. What am i missing here? Thanks!

EDIT: What if I want to ffill within the respective 'ID':

ID     SS     RR
ABC          10.4
ABC   12.6  
ABC     
LMN           5.6
LMN   8.7

Intended result:

ID     SS     RR
ABC          10.4
ABC   12.6   10.4
ABC   12.6   10.4
LMN           5.6
LMN   8.7     5.6

Do i use the groupby function and where to add it to make things work? Thanks again!

Upvotes: 0

Views: 184

Answers (3)

Mayank Porwal
Mayank Porwal

Reputation: 34046

You don't need fillna here, you can call ffill() directly:

In [1793]: df[['SS', 'RR']] = df[['SS', 'RR']].ffill()

In [1794]: df
Out[1794]: 
     SS    RR
0   NaN  10.4
1  12.6  10.4
2  12.6  10.4
3  12.6   5.6
4   8.7   5.6

Upvotes: 1

Cameron Riddell
Cameron Riddell

Reputation: 13407

fillna does not operate inplace. You'll either need to add inplace=True to your calls to fillna as another answer has indicated.

Alternatively you can rewrite your code to overwrite your dataframe with the corrected output.

df = df.replace("", np.nan).ffill()
  • .replace("", np.nan) will replace all blanks with NaN throughout your entire dataframe
  • .ffill() is a shortcut for .fillna(method="ffill")

Since we're not specifying the inplace=True flag here, we can chain these methods together and overwrite our dataframe (or produce a new dataframe with these alterations by changing our variable name:

clean_df = df.replace("", np.nan).ffill()

Upvotes: 1

BENY
BENY

Reputation: 323226

You should assign it back or add inplace

df.SS.fillna(method='ffill',inplace=True)
df.RR.fillna(method='ffill',inplace=True)

#df.SS = df.SS.fillna(method='ffill')
#df.RR = df.RR.fillna(method='ffill')

Upvotes: 2

Related Questions