Append numpy array with diff dimensions

Question

How can I do this? I just need a numpy array/dictionary of the form

[ z (an int), [lamda (np float array)], [flux (np float array)] ]

So I can create the 'data25' numpy array below... But when I try to append another record with the same format, I get an error:

data25 = [0.0,sData1['lambda'],sData1['flux']]
print data25
for z in np.arange(0.2,1.0,0.2):
    newLambdaArr = []
    newFluxArr   = []
    for entry in sData1:
        newLambdaArr = np.append(newLambdaArr, [entry[0] * (1.0 + z)])
        newFluxArr   = np.append(newFluxArr,   [entry[1]/(fpi * frw.dl(z) * (1+z))])
    print 'new lambda',newLambdaArr
    print 'new flux', newFluxArr
    temp = [z, newLambdaArr, newFluxArr]
    print 'just built: ', temp
    data25 =  np.append(data25, [temp] )
    print data25

I get the error at the "data25 = np.append..." line:

ValueError: setting an array element with a sequence.

The first "print data25" gives me what I want:

[0.0, array([  9.10000000e+01,   9.40000000e+01,   9.60000000e+01, ...,
     1.20000000e+06,   1.40000000e+06,   1.60000000e+06], dtype=float32), array([  0.00000000e+00,   0.00000000e+00,   0.00000000e+00, ...,
     3.60900003e-08,   1.95199998e-08,   1.14600001e-08], dtype=float32)]

Thx

Don't use np.append. Use alist.append(newvalue). If you want numpy arrays (as opposed to Python lists), create those explicitly. — hpaulj
– hpaulj, Commented Dec 1, 2014 at 3:23

bakkal · Accepted Answer · 2014-12-01 05:26:47Z

1

numpy.append works differently, it'll return a single list with the elements inside the list you have as argument so it's not what you want:

>>> np.append([1, 2, 3], [4, 5, 6])
array([1, 2, 3, 4, 5, 6])

Your use case can be done with a simple Python list like this:

>>> z = 5
>>> flux = np.array([1, 2, 3])
>>> lambdas = np.array([4, 5, 6])
>>> data = [z, flux, lambdas]
>>> data
[5, array([1, 2, 3]), array([4, 5, 6])]
# Alternatively you can do data.append(z), data.append(flux) and then data.append(lambdas)

If you want the result to be a numpy array, you'll have to shape it like a multidimensional array:

>>> data = np.array([np.array([z]), flux, lambdas]) # Notice the z integer is passed as an array
>>> data
array([array([5]), array([1, 2, 3]), array([4, 5, 6])], dtype=object)

If you don't shape it like that, you'll get your error:

>>> data = np.array([z, flux, lambdas])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: setting an array element with a sequence.

edited Dec 1, 2014 at 5:26

answered Dec 1, 2014 at 5:18

bakkal

55.6k12 gold badges136 silver badges113 bronze badges

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

1 Comment

earnric Over a year ago

Thanks... That: data = np.array([np.array([z]), flux, lambdas]) works perfectly.

Collectives™ on Stack Overflow

Append numpy array with diff dimensions

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related