Don't understand the Numpy Boolean indexing documentation

user_5362515 0
Viewed 213

I am running a very simple example -

a = np.array([1,2,3,4,5])
mask= np.array([True, False])
a[mask]

This produces IndexError: boolean index did not match indexed array along dimension 0; dimension is 5 but corresponding boolean dimension is 2.

The confusion arises from my understanding (or lack thereof) of this part of Numpy doc: -

If obj.ndim == x.ndim, x[obj] returns a 1-dimensional array filled with the elements of x corresponding to the True values of obj. The search order will be row-major, C-style. If obj has True values at entries that are outside of the bounds of x, then an index error will be raised. If obj is smaller than x it is identical to filling it with False. (emphasis mine)

I thought the mask array would be transformed to [True, False, False, False, False] but seemingly that is not the case. Also, both a and mask have same ndim value, so why would error message say a has dimension 5 but boolean dimension is 2.

What am I missing? How do I interpret the doc?

numpy python
2 Answers

As of version 1.13, mask size must match

https://numpy.org/doc/stable/release/1.13.0-notes.html#boolean-indexing-changes

Boolean indexing changes
Boolean array-likes (such as lists of python bools) are always treated as boolean indexes.

Boolean scalars (including python True) are legal boolean indexes and never treated as integers.

Boolean indexes must match the dimension of the axis that they index.

Boolean indexes used on the lhs of an assignment must match the dimensions of the rhs.

Boolean indexing into scalar arrays return a new 1-d array. This means that array(1)[array(True)] gives array([1]) and not the original array.

In v 1.20, there was a further small correction to require matching shape as well as size.

The line you quote appears to be a left over from an earlier time where the size match was not enforced.

Most often the boolean mask is constructed by doing some comparison on the array itself, so the size match is automatic. People don't normally construct short mask - except by mistake or to test the documentation.

user_9019250

As mentioned in the comments, the use of dimension in doc is not consistent, probably due to deprecated content in the old version.

You might regard boolean index as "take the element at the position(True) or not(False)", so the shape of obj and x shall be the same, as mentioned in doc

user_187671600
Related