Output a single row in pandas to an array

Question

I would like to be able to pull out an arbitrary row, such as the 4th row, and get an array so that I can send it through another function.

What would be the easiest way to do this?

Solution 1

# Dummy DataFrame
df = pd.DataFrame({'col1': [1,2,3], 'col2': [4,5,6]})
# Extract second row (index: 1)
df.iloc[1].values

Solution 2

Since Pandas data is stored internally as Numpy arrays, you can extract the Numpy representation directly.

v0.24+

Use pd.Series.to_numpy method:

df.iloc[3].to_numpy()  # output 4th row as Numpy array

Before v0.24

Use pd.Series.values property:

df.iloc[3].values  # output 4th row as Numpy array

Just remember that if your dataframe contains multiple types, the dtype of the row-wise Numpy array may become object. This will lead to inefficiencies in subsequent operations.

Comments

[GoPackGo]

as_matrix was deprecated since version 0.23.0. Generally, it is recommended to use ‘.values’ as shown below in the accepted answer.