CODEWITHSUNDEEP
python-3.xpandas
Nordle
Nordle

Reputation: 2981

Pandas syntax for multiple if statements using | and & operators

If I want to use the following syntax in Pandas to create an if condition using comparison operators, do I need to wrap the or conditions within their own parenthesis?

Without extra parenthesis

df.loc[(df['Col1'] == 'Yes') &
       (df['Col2'] == 'Yes') &
       (df['Col3'] == 'No') |
       (df['Col3'] == 'Yes') |
       (df['Col3'] == 'Maybe'),
       ['Result']] = 'CORRECT'

With extra parenthesis

df.loc[(df['Col1'] == 'Yes') &
       (df['Col2'] == 'Yes') &
       ((df['Col3'] == 'No') |
       (df['Col3'] == 'Yes') |
       (df['Col3'] == 'Maybe')),
       ['Result']] = 'CORRECT'

Or are both valid? It's hard to deduce from the documentation which should be correct.

Upvotes: 1

Views: 208

Answers (1)

jezrael
jezrael

Reputation: 862761

I think it is better to use isin:

df.loc[(df['Col1'] == 'Yes') &
       (df['Col2'] == 'Yes') &
       (df['Col3'].isin(['No', 'Yes', 'Maybe'])), 'Result'] = 'CORRECT'

But your statements are different:

@DeepSpace - comment:

It has nothing to do directly with pandas, but with operator precedence in Python. The bitwise and (&) has precedence over bitwise or (|) so in this case the parenthesis are required. 

np.random.seed(1245)

a = ['No', 'Yes', 'Maybe']
df = pd.DataFrame(np.random.choice(a, size=(10, 3)), columns=['Col1','Col2','Col3'])

df.loc[(df['Col1'] == 'Yes') &
       (df['Col2'] == 'Yes') &
       (df['Col3'] == 'No') |
       (df['Col3'] == 'Yes') |
       (df['Col3'] == 'Maybe'),
       'Result1'] = 'CORRECT'

df.loc[(df['Col1'] == 'Yes') &
       (df['Col2'] == 'Yes') &
       ((df['Col3'] == 'No') |
       (df['Col3'] == 'Yes') |
       (df['Col3'] == 'Maybe')),
       'Result2'] = 'CORRECT'

df.loc[(df['Col1'] == 'Yes') &
       (df['Col2'] == 'Yes') &
       (df['Col3'].isin(['No', 'Yes', 'Maybe'])), 'Result3'] = 'CORRECT'

print (df)
    Col1   Col2   Col3  Result1  Result2  Result3
0  Maybe    Yes    Yes  CORRECT      NaN      NaN
1     No  Maybe     No      NaN      NaN      NaN
2    Yes     No  Maybe  CORRECT      NaN      NaN
3  Maybe    Yes     No      NaN      NaN      NaN
4    Yes    Yes  Maybe  CORRECT  CORRECT  CORRECT
5    Yes  Maybe    Yes  CORRECT      NaN      NaN
6  Maybe  Maybe    Yes  CORRECT      NaN      NaN
7  Maybe     No    Yes  CORRECT      NaN      NaN
8  Maybe  Maybe  Maybe  CORRECT      NaN      NaN
9    Yes    Yes     No  CORRECT  CORRECT  CORRECT

Upvotes: 2

Related Questions