Starting from a df like that:

I want to create a new column comparing two different columns, for do this I defined a function like this:

def F1(k):
    if (k['Lev'] == '.1'):
      return 'Expl1'
    elif (k['Lev'] == '.2'):
      return 'Expl2'
    elif (k['Qtà'] == k['Qty']):
        return 'OK'
    elif (k['Qtà'] > k['Qty']):
        return 'NOK'
    elif (k['Qtà'] < k['Qty']):
        return 'Less'
    elif (k['Qty'] is empty):
        return'NoMap'
    else
        Return 'Check'
df['OK/NOK'] = df.apply(F1, axis=1)

The function works well except when in QTY I have an empty cell. When the cell is empty i receive 'Check' answer.

This is an example of what would like to obtain:

CodePudding user response：

I think it's the "is empty" test, without more context it's hard to tell:

def F1(k):
if k['Lev'] == '.1':
  return 'Expl1'
elif k['Lev'] == '.2':
  return 'Expl2'
elif k['Qtà'] == k['Qty']:
    return 'OK'
elif k['Qtà'] > k['Qty']:
    return 'NOK'
elif k['Qtà'] < k['Qty']:
    return 'Less'
elif not k['Qty']:
    return 'NoMap'
else
    Return 'Check'
df['OK/NOK'] = df.apply(F1, axis=1)

CodePudding user response：

You could use a vectorial approach.

A possibility is to use numpy.select with all individual conditions.

import numpy as np

df['OK/NOK'] = np.select([df['Lev'] == '.1', df['Lev'] == '.2', ...],
                         ['Expl1', 'Expl2', ...],
                         'Check')

Another one would be to use a pandas approach:

import numpy as np

r1 = df['Lev'].map({'.1': 'Expl1', '.2': 'Expl2'})
r2 = np.sign(df['Qta']-df['Qty']).map({1: 'NOK', 0: 'OK', -1: 'LESS'})
r3 = pd.isna(df['Qty']).map({True: 'NoMap', False: 'Check'})

df['OK/NOK'] = r1.fillna(r2).fillna(r3)

output:

  Lev  Qta       Text  Qty OK/NOK
0  .1    1     Model3  1.0  Expl1
1  .2    2    Gruoup1  1.0  Expl2
2  .3    5  Gruoup1.1  4.0    NOK
3  .3    1  Gruoup1.2  1.0     OK
4  .3    2  Gruoup1.3  1.0    NOK
5  .3    1  Gruoup1.4  NaN  NoMap
6  .3    1  Gruoup1.5  NaN  NoMap