Project selected columns
Retain a known subset of columns by passing their exact names.
Pandas · Module
df.filter
DataFrame.filter selects rows or columns by their labels using exact names, substring matching, or regular expressions. It does not inspect the values in the DataFrame.
How it works
DataFrame.filter returns a same-type object containing labels selected by an explicit collection, substring match, or regular expression. For DataFrames, the default axis is columns; set axis to index or 0 to select rows, and remember that matching examines labels rather than cell values.
Syntax
DataFrame.filter(*items=None*, *like=None*, *regex=None*, *axis=None*)Parameters
| Name | Type | Description |
|---|---|---|
items | list-like, optional | Labels to retain when they occur in the selected axis. |
like | str, optional | Substring that must occur in a selected-axis label. |
regex | str, optional | Regular expression searched against each selected-axis label. |
axis | {0 or ‘index’, 1 or ‘columns’, None}, default None | Axis to inspect: 0 or index for rows, 1 or columns for columns, or None. |
Returns
A DataFrame containing the selected labels; the returned object has the same type as the input object.
Key points
- The method creates a subset based on index labels.
- items performs exact membership selection against the chosen axis labels.
- like retains labels containing the supplied substring.
- regex uses regular-expression search to decide which labels remain.
- When axis is omitted for a DataFrame, filtering applies to columns.
Common mistakes
- It selects according to row or column labels, not the values stored in the DataFrame.
Example: Filter labels on either axis
import pandas as pd
df = pd.DataFrame(
[[1, 2, 3], [4, 5, 6]],
index=["mouse", "rabbit"],
columns=["one", "two", "three"],
)
columns = df.filter(items=["one", "three"])
rows = df.filter(like="bbi", axis="index")
regex_columns = df.filter(regex="e$", axis="columns")
print(columns)
print(rows)
print(regex_columns)Selects named columns, an index label containing a substring, and columns whose labels end in “e”.
When to use it
Locate labeled records
Find rows whose index labels contain a meaningful substring.
Match labels with regex
Select columns following a naming convention, such as a suffix or prefix.
Interview questions
QWhat does DataFrame.filter do?
DataFrame.filter selects a subset using labels. The available modes are exact labels through items, substring matching through like, or regular expressions through regex; the method returns the same general type as its input.
Follow-up: How would you select rows rather than columns? Pass axis="index" or axis=0, because a DataFrame defaults to filtering its columns.
QHow does axis affect filtering?
The default axis is the DataFrame’s information axis, which is columns. Set axis="index" or axis=0 when matching row labels, or axis="columns" or axis=1 when matching column labels.
Follow-up: What happens if axis is omitted while using like? The substring is matched against column labels.
QHow do the three selection modes differ?
items keeps labels present in the supplied list-like object, like keeps labels containing a substring, and regex keeps labels for which a regular-expression search succeeds. These modes operate on labels, not data values.
Follow-up: How can columns ending in “e” be selected? Use regex="e$" with axis=1 or axis="columns".
Related
Series.filterDataFrame.locDataFrame.select_dtypes