Python creating a smaller sub-array from a larger 2D NumPy array?

Question

So I have a large NumPy array that takes the following form:

data = [[2456447.64798471, 4, 15.717, 0.007, 5, 17.308, 0.019, 6, 13.965, 0.006],
        [2456447.6482855, 4, 15.768, 0.018, 5, 17.347, 0.024, 6, 14.001, 0.023],
        [2456447.648575, 4, 15.824, 0.02, 5, 17.383, 0.024, 6, 14.055, 0.023]]

I want to create a sub array that looks like this:

[[4, 15.717, 5, 17.308, 6, 13.965], 
 [4, 15.768, 5, 17.347, 6, 14.001],
 [4, 15.824, 5, 17.383, 6, 14.055]]

Basically I want to select out the first column, and then starting at the 4th column I want to select out every 3rd column. I tried to figure this out how to approach this with something like data[1:6:?] but I didn't understand how to step through and only get the columns that I wanted.

Also I need this to be scalable for an array that extends horizontally. So I don't just want to hard code the column values.

Eric Workman · Accepted Answer · 2013-06-28 17:16:43Z

7

You could do this:

>>> data[:, [1, 2, 4, 5, 7, 8]]
array([[  4.   ,  15.717,   5.   ,  17.308,   6.   ,  13.965],
       [  4.   ,  15.768,   5.   ,  17.347,   6.   ,  14.001],
       [  4.   ,  15.824,   5.   ,  17.383,   6.   ,  14.055]])

answered Jun 28, 2013 at 17:16

Eric Workman

1,45512 silver badges13 bronze badges

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

4 Comments

Stephan Over a year ago

what is going on here, slice notation with a list? and why is there a comma after the colon?

Eric Workman Over a year ago

@Stephan, since data is a NumPy array, it can be accessed with integer indexing. The list can be another array too. The docs for it are here. The comma separates slicing along the axes. So data[1:2, :] would select the first row and all the columns, while data[:, 1:2] would select all the rows and the first column.

Stephan Over a year ago

are you sure data[1:2] doesn't get the second item? I thought slice was 0 indexed so data[0:1] would get the first item

Eric Workman Over a year ago

First as in directly after zeroth. I was following the notation in the original question. Sorry I wasn't clear.

Stephan · Accepted Answer · 2013-06-28 18:31:56Z

3

This will do the trick, it scales horizontally and vertically and it's easy and works.

subArray = []
newRow = []
for row in data:
    for i in xrange(0,len(row)):
        if (i % 3 == 0):
            continue
        newRow.append(row[i])
    subArray.append(newRow)
    newRow = []

edited Jun 28, 2013 at 18:31

answered Jun 28, 2013 at 17:05

Stephan

18.4k8 gold badges38 silver badges61 bronze badges

5 Comments

Stephan Over a year ago

if you need this to be scalable and not statically sized let me knkow and I can change the code

sTr8_Struggin Over a year ago

Yeah I was trying to make something scalable because my actual data array is much longer both horizontally and vertically. If you have a scalable solution that would be awesome!

sTr8_Struggin Over a year ago

Thanks for the help so far. I ran your code and I got the following error: if (row.index(element) % 3 == 0): AttributeError: 'numpy.ndarray' object has no attribute 'index'

Stephan Over a year ago

@sTr8_Struggin that's because your data is actually a numpy array, try my new code

Stephan Over a year ago

@sTr8_Struggin If you are using python 3.x you should take the x off of xrange btw

Collectives™ on Stack Overflow

Python creating a smaller sub-array from a larger 2D NumPy array?

2 Answers 2

4 Comments

5 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

5 Comments

Your Answer

Sign up or log in

Post as a guest

Related