Fill NaN values of selected dataframe columns using For loop-CodePudding

I have a dataframe of patients Vital signs (HR, O2Sat, Temp, SBP, DBP, Resp) with NaN values. I filled NaN in individual patient based on Patient ID (P_ID) column using the code:

m['HR'] = m['HR'].fillna(m.groupby('P_ID')['HR'].transform('mean'))
m['O2Sat'] = m['O2Sat'].fillna(m.groupby('P_ID')['O2Sat'].transform('mean'))
m['Temp'] = m['Temp'].fillna(m.groupby('P_ID')['Temp'].transform('mean'))
m['SBP'] = m['SBP'].fillna(m.groupby('P_ID')['SBP'].transform('mean'))
m['DBP'] = m['DBP'].fillna(m.groupby('P_ID')['DBP'].transform('mean'))
m['Resp'] = m['Resp'].fillna(m.groupby('P_ID')['Resp'].transform('mean'))

It worked perfectly. However it is a lot of code. Is there anyway I use for loop to fill NaN values in only the vital columns? As there are some more columns without NaN values. Thanks.

CodePudding user response：

Yes, you can use a loop

for i in ['HR','O2Sat','Temp','SBP','DBP','Resp']:
    m[i] = m[i].fillna(m.groupby('P_ID')[i].transform('mean'))

CodePudding user response：

In the interest of avoiding unnecessary lines and lengthy scripts, you can use the columns attribute of the dataframe. But note that, since you mentioned vital columns, there could be a few columns that you don't want to do the following code.:

for column in m.columns:
  m[column] = m.fillna(m.groupby('P_ID')[column].transform('mean'))