CODEWITHSUNDEEP
pythonarrayspython-2.7numpy
Luna Álvarez Jiménez
Luna Álvarez Jiménez

Reputation: 33

Split an array according to it's information

I'm doing a project in python which has an array of 600 asteroid measurements (diameter, period, orbit radius). With that information I'd like to make new arrays according the diameter column, the resulting arrays would be diameter intervals of the first one.

If i have this array, 

diameter           period         orbit radius
 [[  19.17          5.71476129    3.19639121]
 [  19.28          4.81234455    2.85035536]
 [  22.77          5.62890294    3.16429553]
 [   4.8           3.29145453    2.21268592]
 [   7.23          2.61331495    1.89724041]
 [   8.17          2.54935585    1.86615697]
 [ 260.94          6.49151957    3.47983602]
 [ 530.            3.62867648    2.3613482 ]
 [ 952.4           4.60562864    2.76813421]]

I'd like to make an array with diameter from 0 to 20

diameter           period         orbit radius
 [[  19.17          5.71476129    3.19639121]
 [  19.28          4.81234455    2.85035536]
 [   4.8           3.29145453    2.21268592]
 [   7.23          2.61331495    1.89724041]
 [   8.17          2.54935585    1.86615697]]

Upvotes: 3

Views: 100

Answers (3)

Nickil Maveli
Nickil Maveli

Reputation: 29711

Assuming you would like to achieve the required matrix using numpy, you may use numpy.where to match the conditions and then apply a mask which can be used to index the array to get the required values.

>>>arr[np.where(arr[:, 0] < 20)]

[[ 19.17         5.71476129   3.19639121]
 [ 19.28         4.81234455   2.85035536]
 [  4.8          3.29145453   2.21268592]
 [  7.23         2.61331495   1.89724041]
 [  8.17         2.54935585   1.86615697]]

Here, np.where(arr[:, 0] < 20) checks first column of the array and returns the indices of rows where there's a favourable match which would be used to slice the original array later.

Upvotes: 1

jadsq
jadsq

Reputation: 3382

You can easily build new array using the indexing properties of numpy arrays, and that translates into that one liner :

filtered_arr = arr[(arr[:,0]<=20)*(arr[:,0]>=0),:]

or to develop:

#we separate the diameters from the rest
diameter_column= arr[:,0]
#we create a (1D) bool array that match diameter between 0 and 20
diameter_filter=(diameter_column<=20) * (diameter_column>=0)
#we use the boolean array as an index to select the right asteroids
filtered_arr=arr[diameter_filter,:]

Upvotes: 1

Brian
Brian

Reputation: 1667

Just use a list comprehension to filter it

filtered = [ ast for ast in arr if ast[0]<=20 ]

Upvotes: 4

Related Questions