treat list of numpy arrays as a numpy object

Question

I am using a data object that is a list of numpy arrays of different shapes. I would like to treat the data object as a single numpy object, such that I can add, subtract and scale these objects as if they were pure numpy arrays:

# Defining the data objects:
d1 = [numpy.random.rand(n**2).reshape(n, n) for n in range(1, 5)]
d2 = [numpy.random.rand(n**2).reshape(n, n) for n in range(1, 5)]

# The operation I want to perform:
d3 = [a1 + 2*a2 for a1, a2 in zip(d1, d2)]

# What I really would like to write:
d3 = d1 + 2*d2   # (not valid for lists of arrays)

What is the easiest way to define the objects d1 and d2 such that I can use the last line?

The reason I ask is because I want to reuse some code that assumes that d1 and d2 are simple numpy arrays that can be added and scaled.

@Moses Koledoye: The shapes of the arrays match. See the example. — Quant Metropolis
– Quant Metropolis, Commented Jan 2, 2017 at 14:28
Pay attention to shape and dtype when creating an array from these lists. Since the subarrays differ in shape, the combination will be a 1d array of array objects, not a multidimensional array of floats. Some, but not all, numpy math propagates through those objects. — hpaulj
– hpaulj, Commented Jan 2, 2017 at 17:32

glS · Accepted Answer · 2017-01-02 14:30:39Z

1

Simply defining them as numpy.array works:

d1 = numpy.array([numpy.random.rand(n**2).reshape(n, n) for n in range(1, 5)])
d2 = numpy.array([numpy.random.rand(n**2).reshape(n, n) for n in range(1, 5)])
d3 = d1 + 2 * d2

Or if you want to keep d1 and d2 as regular lists you can use numpy.asarray:

d1_ = numpy.asarray(d1)
d2_ = numpy.asarray(d2)
d3 = d1_ + 2 * d2_

answered Jan 2, 2017 at 14:30

glS

1,2997 gold badges19 silver badges51 bronze badges

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

2 Comments

hpaulj Over a year ago

What does asarray do that's different?

glS Over a year ago

@hpaulj nothing. I just meant it as an option if one already has the lists stored somewhere and does not want to change the defining lines

Collectives™ on Stack Overflow

treat list of numpy arrays as a numpy object

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related