Numpy, new array based on an array within an array

Question

I'm looping over an array to extract data but it's too slow. Is there a function in numpy that will do this faster? I have to create a new array based on values from an array within an array.

For example, create an array with cars made in USA.

    input: array = [['ford', 'USA'], ['volkswagen', 'Germany'], ['jeep', 'USA']]

    output: new_array = [['ford', 'USA'], ['jeep', 'USA']]

Divakar · Answer 1 · 2016-10-28T13:09:33.237

1

Assuming an array of string dtype, you can slice out the second column and compare against the string 'USA' to get a boolean array. This boolean array could then be used for indexing into the array using boolean-indexing to select valid rows and give us the desired output.

Thus, the implementation would be simply -

array[array[:,1] == 'USA']

edited Oct 28 '16 at 13:09

answered Oct 28 '16 at 13:05

Divakar

218,885
19
262
358

Thank you! @Divakar – elfving Oct 28 '16 at 13:09
How would I write it if I want a result based on two values? For exampel 'jeep' and 'USA' – elfving Oct 29 '16 at 10:22
@elfving Use `np.in1d`. – Divakar Oct 29 '16 at 10:27
@elfving Also have a look at [this Q&A](http://stackoverflow.com/questions/38674027/find-the-row-indexes-of-several-values-in-a-numpy-array) to get those indices and then simply index into the list with those indices. – Divakar Oct 29 '16 at 19:41

Numpy, new array based on an array within an array

1 Answers1