How to Convert Variable-Length Python Lists to a Dense NumPy Array with Placeholders?

Efficiently Convert Python Variable-Length Lists to a Dense NumPy Array

The direct conversion of variable-length Python lists into a NumPy array results in an array of type "object," which can be undesirable. Alternatively, attempting to force a specific type using np.array(v, dtype=np.int32) leads to an exception due to the presence of sequences in the array.

Therefore, to create a dense NumPy array of a specific data type (e.g., int32) while filling missing values with a placeholder, you can leverage the itertools.zip_longest function.

For example, considering the input sequence v = [[1], [1, 2]], using itertools.zip_longest with a placeholder of 0, you can efficiently obtain a dense NumPy array as follows:

<code class="python">import itertools
np.array(list(itertools.zip_longest(*v, fillvalue=0))).T</code>

This will produce the desired output:

array([[1, 0],
       [1, 2]])

Note that for Python 2, use itertools.izip_longest instead.

The above is the detailed content of How to Convert Variable-Length Python Lists to a Dense NumPy Array with Placeholders?. For more information, please follow other related articles on the PHP Chinese website!