Here’s my quick cheat-sheet on slicing columns from a Pandas dataframe. Consider this dataset:
df = pd.DataFrame([
["AA", "747", 1],
["AB", "A380", 1],
["AA", "737", 0],
["AB", "747", 1],
["AA", "737", 0]
],
columns=["Airline", "Aircraft", "Class"])
It has 3 columns. You can use the iloc function to slice columns:
| Airline | Aircraft | Class |
| AA | 747 | 1 |
| AB | A380 | 1 |
| AA | 737 | 0 |
| AB | 747 | 1 |
| AA | 737 | 0 |
1st column only:
df.iloc[:, 0:1].values
array([['AA', '747'],
['AB', 'A380'],
['AA', '737'],
['AB', '747'],
['AA', '737']], dtype=object)
1st through last columns (i.e. all columns):
df.iloc[:, :].values
df.iloc[:, 0:].values # Alternate syntax
array([['AA', '747', 1],
['AB', 'A380', 1],
['AA', '737', 0],
['AB', '747', 1],
['AA', '737', 0]], dtype=object)
2nd column only:
df.iloc[:, 1:2].values
array([['747'],
['A380'],
['737'],
['747'],
['737']], dtype=object)
2nd through last columns:
df.iloc[:, 1:].values
array([['747', 1],
['A380', 1],
['737', 0],
['747', 1],
['737', 0]], dtype=object)
1st through last-but-one:
df.iloc[:, :-1].values
df.iloc[:, 0:-1].values # Alternate syntax
array([['AA', '747'],
['AB', 'A380'],
['AA', '737'],
['AB', '747'],
['AA', '737']], dtype=object)
Last column only:
df.iloc[:, -1:].values
array([[1],
[1],
[0],
[1],
[0]])