Let's say I have the following dataframe:
| index | A |
|---|---|
| 1 | 3 |
| 2 | 5 |
| 3 | 20 |
| 4 | 8 |
| 5 | 7 |
| 6 | 13 |
| 7 | 33 |
| 8 | 2 |
I want to create a new column which is based on column A and create groups depending on the value of column A.
df = pd.DataFrame([['1', 3],['2', 5],['3', 20],['4',8],['5',7],['6',13],['7',33],['8',2]], columns=['index', 'A'])
df['groups'] = df['A'].apply(lambda x: 'high' if x>15 else 'medium' if 15>=x>10 else 'low')
How could I do the same using assign?
df = df\
.assign(groups = ?)
You could use np.select here for a much more efficient approach for cases with multiple conditions. assign is just a method to add a new column, which could also just be done with panda's indexing methods.
import numpy as np
df.assign(groups=np.select([df.A>15, (df.A<=15)&(df.A>10)],
['high','medium'], 'low'))
index A groups
0 1 3 low
1 2 5 low
2 3 20 high
3 4 8 low
4 5 7 low
5 6 13 medium
6 7 33 high
7 8 2 low
Use assign() method:
df=df.assign(groups =df['A'].apply(lambda x: 'high' if x>15 else 'medium' if 15>=x>10 else 'low'))
OR
value=df['A'].apply(lambda x: 'high' if x>15 else 'medium' if 15>=x>10 else 'low')
df=df.assign(groups=value)
Now if you print your df you will get your desired output:
#output
index A groups
0 1 3 low
1 2 5 low
2 3 20 high
3 4 8 low
4 5 7 low
5 6 13 medium
6 7 33 high
7 8 2 low
Edit: you can also do this by:
result=pd.read_csv(file).assign(groups=pd.read_csv(file)['A'].apply(lambda x: 'high' if x>15 else 'medium' if 15>=x>10 else 'low'))
pd.cut(df['A'], bins=(-np.inf, 10, 15, np.inf), labels=['low','med','high'])