What's the Difference Between Pandas' `iloc` and `loc` for Data Selection?

How iloc and loc Differ: Label vs. Location

Understanding the Distinction

The primary distinction between iloc and loc lies in how they access rows and columns:

• loc: Locates data using row and column labels. These labels are typically index values or column names.
• iloc: Locates data using row and column integer locations. These locations refer to the position of the elements in the DataFrame.

Demonstration

Consider the example DataFrame below:

Index	Column A
0	John
1	Mary
2	Peter

Extracting the first 5 rows:

• loc[:5]: Returns all rows with index labels 0 to 4 (inclusive).
• iloc[:5]: Returns the first 5 rows at integer locations 0 to 4 (exclusive).

Clarifying the Difference

To further illustrate, consider a non-monotonic index:

Index	Series
49	a
48	b
47	c
0	d
1	e
2	f

Accessing the value at index label 0:

• loc[0] fetches 'd' because it uses index labels.
• iloc[0] fetches 'a' because it uses integer locations (even though the integer location of 'd' is 3).

Accessing a slice of rows:

• loc[0:1] retrieves rows with index labels 0 and 1 (inclusive).
• iloc[0:1] retrieves only the row at index location 0 (and does not include row 1).

Additional Considerations

• Missing labels: loc raises a KeyError if the specified label is not in the index, while iloc returns an IndexError.
• Boolean Series: loc can index through a Boolean Series, while iloc returns a NotImplementedError.
• Callables: loc and iloc can both apply callables as indexers, but they handle out-of-bounds values differently.

The above is the detailed content of What's the Difference Between Pandas' `iloc` and `loc` for Data Selection?. For more information, please follow other related articles on the PHP Chinese website!