Pandas 删除逗号分隔列值中的特定 int 值

Question

I have a dataframe with column values separated by comma. I want to remove certain values from those values.

My dataframe looks like this:

  col1             col2  
0,1,0,2,30,10,20  0,0,2,3,10,20
0,0,0,1,0,210,30  0,0,20,20,20,0,0,0

I want to remove 0,1,2 from column

Output should be:

 col1             col2                new_col1  new_col2
0,1,0,2,30,10,20  0,0,2,3,10,20       30,10,20   3,10,20
0,0,0,1,0,210,30  0,0,20,20,20,0,0,0   210,30    20,20,20

I tried

def mysub(r):

     lst = [float(a) for a in r.split(',') if a != '0' and a != '' and  a != "1" and  a != "2"]
     return lst
df['new_col1']=df[df['col1']].mysub()

I am not able fix my problem - help me to sort it.

Answer 1

Use list comprehension with specified values for remove in list:

def mysub(r):
    return [','.join(z for z in str(y).split(',') 
            if z not in ['0','1','2']) for y in r]
df = df.apply(mysub)
print (df)
       col1      col2
0  30,10,20   3,10,20
1    210,30  20,20,20

For new columns:

def mysub(r):
    return [','.join(z for z in str(y).split(',') 
            if z not in ['0','1','2']) for y in r]
df = df.join(df.apply(mysub).add_prefix('new_'))
print (df)
               col1                col2  new_col1  new_col2
0  0,1,0,2,30,10,20       0,0,2,3,10,20  30,10,20   3,10,20
1  0,0,0,1,0,210,30  0,0,20,20,20,0,0,0    210,30  20,20,20

If want floats output:

def mysub(r):
    return [[float(z) for z in str(y).split(',') 
            if z not in ['0','1','2']] for y in r]
df = df.join(df.apply(mysub).add_prefix('new_'))
print (df)
               col1                col2            new_col1  \
0  0,1,0,2,30,10,20       0,0,2,3,10,20  [30.0, 10.0, 20.0]   
1  0,0,0,1,0,210,30  0,0,20,20,20,0,0,0       [210.0, 30.0]   

             new_col2  
0   [3.0, 10.0, 20.0]  
1  [20.0, 20.0, 20.0]  

Answer 2

df.applymap(lambda x: ','.join([e for e in x.split(',') if e not in ['0','1','2']]))