Example: Filtering with axis set to 'rows' or 0
When axis is 'rows' or 0, the items list specifies the values of the column(s) specified in the index_label of the DataFrame. Only rows where the values of the index column(s) in the items list are returned.
>>> df = DataFrame('admissions_train', index_label = ['programming'])
>>> df
id masters gpa stats admitted
programming
Advanced 15 yes 4.00 Advanced 1
Beginner 34 yes 3.85 Advanced 0
Novice 13 no 4.00 Advanced 1
Beginner 38 yes 2.65 Advanced 1
Novice 5 no 3.44 Novice 0
Beginner 40 yes 3.95 Novice 0
Novice 7 yes 2.33 Novice 1
Beginner 22 yes 3.46 Novice 0
Advanced 26 yes 3.57 Advanced 1
Advanced 17 no 3.83 Advanced 1
>>> df.filter(items = ['Advanced', 'Novice'], axis = 0)
id masters gpa stats admitted
programming
Advanced 15 yes 4.00 Advanced 1
Novice 37 no 3.52 Novice 1
Novice 12 no 3.65 Novice 1
Advanced 17 no 3.83 Advanced 1
Advanced 11 no 3.13 Advanced 1
Advanced 26 yes 3.57 Advanced 1
Novice 5 no 3.44 Novice 0
Novice 24 no 1.87 Advanced 1
Novice 13 no 4.00 Advanced 1
Novice 7 yes 2.33 Novice 1
Example: Filtering with axis set to 'columns' or 1
When axis is 'columns' or 1, then column names provided in the items list are selected from the DataFrame.
>>> df.filter(items = ['programming', 'id', 'masters', 'gpa'])
id masters gpa
programming
Advanced 15 yes 4.00
Novice 7 yes 2.33
Beginner 22 yes 3.46
Advanced 17 no 3.83
Novice 13 no 4.00
Beginner 38 yes 2.65
Advanced 26 yes 3.57
Novice 5 no 3.44
Beginner 34 yes 3.85
Beginner 40 yes 3.95
Supported types for python literals in items
You can specify float, decimal.Decimal, str, bytes, datetime.time, datetime.date, and datetime.datetime python literal types in the items list.