Home > Enterprise >  AttributeError: 'DataFrame' object has no attribute 'assign'
AttributeError: 'DataFrame' object has no attribute 'assign'

Time:10-27

While assign a new column to my DataFrame I get this error here is my code

def check_header(header, df):
    print("Header : ",header)
    for item in header:
        if not item in df.columns:
            df = df.assign(item) #here I'm getting error
    return df[header]

I have check this post but din't work for me because my pandas version is satisfying it's

>>> import pandas as pd
>>> pd.__version__
'1.1.5'

what is problem in my code please help me.

CodePudding user response:

If need add new columns by header list with convert not matched values to new columns filled by NaNs use DataFrame.reindex:

df = pd.DataFrame(data = {"test":["mkt1","mkt2","mkt3"],
                         "test2":["cty1","cty2","cty3"]})


def check_header(header, df):
    return df.reindex(header, axis=1)

a = ['test','test1','test3']
print (check_header(a, df))
   test  test1  test3
0  mkt1    NaN    NaN
1  mkt2    NaN    NaN
2  mkt3    NaN    NaN

If need same values in new columns use fill_value parameter:

def check_header(header, df):
    return df.reindex(header, axis=1, fill_value=0)

a = ['test','test1','test3']
print (check_header(a, df))
   test  test1  test3
0  mkt1      0      0
1  mkt2      0      0
2  mkt3      0      0

If need different values per new column use DataFrame.assign with dictionary for new columns names like keys::

def check_header(header, df):
    diff = np.setdiff1d(header, df.columns)
    d = dict(zip(diff, diff))
    print (d)
    {'test1': 'test1', 'test3': 'test3'}
    
    return df.assign(**d).reindex(header, axis=1)

a = ['test','test1','test3']
print (check_header(a, df))
   test  test1  test3
0  mkt1  test1  test3
1  mkt2  test1  test3
2  mkt3  test1  test3
  • Related