change data type of a array in python

Question

I have created a array by generating a group of random numbers and converted them into int type. However, i do not think my way below is efficient. Is there a best way to change the data type in a array?

# standard normal distributed random numbers
c=random.randn(5,5)
c

array([[-0.37644781, -0.81347483, -0.36895952, -2.68702544, -0.96780752],
   [ 0.05293328,  1.65260753,  0.55586611, -0.5786392 ,  0.50865003],
   [ 1.25508358,  0.51783276,  2.36435212, -0.23484705, -1.20999296],
   [ 2.07552172,  0.65506648,  0.10231436, -0.26046045,  0.40148111],
   [ 0.24864496, -1.8852587 , -2.51091886,  1.01106003,  1.53118353]])

d=array([[-0.37644781, -0.81347483, -0.36895952, -2.68702544, -0.96780752],
       [ 0.05293328,  1.65260753,  0.55586611, -0.5786392 ,  0.50865003],
       [ 1.25508358,  0.51783276,  2.36435212, -0.23484705, -1.20999296],
       [ 2.07552172,  0.65506648,  0.10231436, -0.26046045,  0.40148111],
       [ 0.24864496, -1.8852587 , -2.51091886,  1.01106003,  1.53118353]],dtype=int)

d
array([[ 0,  0,  0, -2,  0],
       [ 0,  1,  0,  0,  0],
       [ 1,  0,  2,  0, -1],
       [ 2,  0,  0,  0,  0],
       [ 0, -1, -2,  1,  1]])

Do you need them normally distributed? If not, np.random.random_integers(-2, 2, (5, 5)) or similar might do it. — jonrsharpe
– jonrsharpe, Commented Apr 24, 2015 at 10:46
I understand what you say, but my object is to create a array by using any way and change the data type in the array to another type,such as from int to float or from int to complex. — Michael Ji
– Michael Ji, Commented Apr 24, 2015 at 10:52

plonser · Accepted Answer · 2015-04-24 10:51:07Z

13

Either you find a way to get the output with the correct type or you use astype, see the docs, in order to change the type of an array

In your case the following example gives you an array of type np.int

c=random.randn(5,5).astype(np.int)

edited Apr 24, 2015 at 10:51

answered Apr 24, 2015 at 10:48

plonser

3,3632 gold badges21 silver badges23 bronze badges

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

5 Comments

jonrsharpe Over a year ago

Would this actually be more efficient? It performs the same steps, creating a float array then converting it to integer.

plonser Over a year ago

@jonrsharpe : I don't think so ... but you avoid copying the output into a new array by hand

abarnert Over a year ago

@jonrsharpe: Even though the OP asked for "efficient", I suspect he doesn't really mean "efficient". You really think the cost of copying-and-casting a 5x5 NumPy array is a bottleneck in his program? Or that he finds manually printing out an array, copying and pasting and editing it is an inefficient way to code?

jonrsharpe Over a year ago

@abarnert frankly, I assumed that: 1. they meant what they said; and 2. this was just a small example to show the process. Who knows, though.

abarnert Over a year ago

@jonrsharpe: I've gotten used to novices using the word "efficient" to mean… well, I'm never sure what, it's always vague, but "fast enough performance for my use case" or "algorithmically optimal performance" or anything remotely similar is rarely even close to whatever they're after…

Ben · Accepted Answer · 2015-04-24 12:24:34Z

2

Also: np.cast[int](c), which doesn't presume it's argument is a NumPy array.

answered Apr 24, 2015 at 12:24

Ben

9,8933 gold badges50 silver badges62 bronze badges

Collectives™ on Stack Overflow

change data type of a array in python

2 Answers 2

5 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

5 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related