Item assignment to Python dask array objects

Question

I've created a Python dask array and I'm trying to modify a slice of the array as follows:

import numpy as np
import dask.array as da

x = np.random.random((20000, 100, 100)) # Create numpy array
dx = da.from_array(x, chunks=(x.shape[0], 10, 10)) # Create dask array from numpy array

dx[:50, :, :] = 0 # Modify a slice of the dask array

Such an attempt to modify the dask array raises the exception:

TypeError: 'Array' object does not support item assignment

Is there a way to modify a dask array slice without raising an exception?

MRocklin · Accepted Answer · 2016-03-21 23:01:21Z

5

Currently dask.array does not support item assignment or any other mutation operation.

In the case above I recommend concatenating with zeros

In [1]: import dask.array as da

In [2]: dx = da.random.random((20000 - 50, 100, 100), chunks=(None, 10, 10))

In [3]: z = da.zeros((50, 100, 100), chunks=(50, 10, 10))

In [4]: dx2 = da.concatenate([z, dx], axis=0)

In [5]: dx2
Out[5]: dask.array<concate..., shape=(20000, 100, 100), dtype=float64, chunksize=(50, 10, 10)>

In [6]: (dx2 == 0)[0:100, 0, 0].compute()
Out[6]: 
array([ True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True,  True,  True,  True,  True,
        True,  True,  True,  True,  True, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False, False], dtype=bool)

The da.where(condition, iftrue, iffalse) function can also be quite useful in working around cases where mutation is often desired.

answered Mar 21, 2016 at 23:01

MRocklin

57.5k29 gold badges176 silver badges245 bronze badges

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

2 Comments

Lcg3 Over a year ago

Thank you, MRocklin! This works great! Now consider a case where you have a list of values (e.g., index = [10, 20, 25, 50, 100, 120]) to index the 0th axis of dx, and all elements corresponding to these index values need to be assigned 0. Concatenation wouldn't work in this case, and as far as I can tell, da.where wouldn't work either. Do you know a workaround for this particular case?

jadelord Over a year ago

Recently dask array mutation (or __setitem__) became supported and the original code just works. See github.com/dask/dask/issues/2000 for some history.

Collectives™ on Stack Overflow

Item assignment to Python dask array objects

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related