Removing "rows" containing negative values in 3D numpy array

Question

What I have:

>>> forces
array([[[  63.82078252,    0.63841691],
        [ -62.45826693,    7.11946976],
        [ -87.85946925,   15.1988562 ],
        [-120.49417797,  -16.31785819],
        [ -81.36080338,    6.45074645]],

       [[ 364.99959095,    4.92473888],
        [ 236.5762723 ,   -7.22959548],
        [  69.55657789,    1.20164815],
        [ -22.1684177 ,   13.42611095],
        [ -91.19739147,  -16.15076634]]])

forces[0] and forces [1] each contain a list of paired values, e.g. 63.82078252 & 0.63841691 are one data point.

I want to remove all pairs where the first value is negative:

>>> forces
array([[[  63.82078252,    0.63841691]],

       [[ 364.99959095,    4.92473888],
        [ 236.5762723 ,   -7.22959548],
        [  69.55657789,    1.20164815]]])

But this type of structure is not possible since the two slices of forces have different sizes: (1, 2) and (3, 2) respectively.

My sloppy attempt:

>>> forces[:,:,0][forces[:,:,0] < 0] = np.nan
>>> forces
array([[[  63.82078252,    0.63841691],
        [          nan,    7.11946976],
        [          nan,   15.1988562 ],
        [          nan,  -16.31785819],
        [          nan,    6.45074645]],

       [[ 364.99959095,    4.92473888],
        [ 236.5762723 ,   -7.22959548],
        [  69.55657789,    1.20164815],
        [          nan,   13.42611095],
        [          nan,  -16.15076634]]])

and then using isnan to remove the relevant entries:

>>> forces = forces[~np.isnan(forces).any(axis=2)]
>>> forces
array([[  63.82078252,    0.63841691],
       [ 364.99959095,    4.92473888],
       [ 236.5762723 ,   -7.22959548],
       [  69.55657789,    1.20164815]])

So these are the correct values but they are now lumped together into a 2D array.

How can I create a heterogeneously sized "array" that will contain two slices of size (1, 2) and (3, 2) respectively, for this simplified example?

Also any pointers on accomplishing the task more elegantly would be much appreciated!

YXD · Accepted Answer · 2013-10-02 15:28:12Z

6

It is simply

forces[forces[..., 0] >= 0]

Read more here: http://scipy-lectures.github.io/intro/numpy/array_object.html#fancy-indexing

answered Oct 2, 2013 at 15:28

YXD

32.6k15 gold badges79 silver badges117 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

CTKlein Over a year ago

That is really good to know (and certainly much cleaner) but it still returns a 2D array.

YXD Over a year ago

I'm not clear on what what shape do you expect to get. forces.shape is (2, 5, 2), so each slice forces[i, ...] has shape (5, 2), i.e. contains five paired values. In your example, filtering slice forces[0, ...] leaves us with one paired set of values and filtering slice forces[1, ...] leaves us with three paired values. How do they sit together in a 3D array?

CTKlein Over a year ago

I think I see my problem. What I was envisioning was forces having 2 slices of shape (1, 2) and (3, 2) respectively which isn't possible in numpy. Any idea on how to achieve a similar result?

YXD Over a year ago

That depends on what your data represents. Are forces[0, 0, :] and forces[1, 0, :], i.e. array([ 63.82078252, 0.63841691]) and array([ 364.99959095, 4.92473888]) supposed to be linked in some way? What are you modelling? Why are there two sets of forces?

CTKlein Over a year ago

One slice, e.g. forces[0], contains a set of x- and y-directed force values at different intervals. forces[1] contains equivalent data for a different system. I simplified the example but I have many more systems and want to plot each system as its own line on a plot but only the points with x-values larger than zero.

|

Collectives™ on Stack Overflow

Removing "rows" containing negative values in 3D numpy array

What I have:

I want to remove all pairs where the first value is negative:

My sloppy attempt:

How can I create a heterogeneously sized "array" that will contain two slices of size (1, 2) and (3, 2) respectively, for this simplified example?

1 Answer 1

6 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

What I have:

I want to remove all pairs where the first value is negative:

My sloppy attempt:

How can I create a heterogeneously sized "array" that will contain two slices of size (1, 2) and (3, 2) respectively, for this simplified example?

1 Answer 1

6 Comments

Your Answer

Sign up or log in

Post as a guest

Related