I would like to concatenate all numpy arrays (all same size) in a python dict to a new array.
dict = {'some_key_1':np.array([1,2,3,4]), 'some_key_2':np.array([2,3,4,5]), ...}
the result that I would like to see is:
result = np.array([[1,2,3,4],[2,3,4,5]])
whats the best way to do that?
Edit: Solution seems to be slightly different for python 2 and 3. The accepted answer solves it for python 3.
In Python 3, you'd need to convert the dict.values() to a list first.
The objects returned by dict.keys(), dict.values() and dict.items() are view objects. They provide a dynamic view on the dictionary’s entries, which means that when the dictionary changes, the view reflects these changes. (https://docs.python.org/3/library/stdtypes.html#typesmapping)
import numpy as np
data = {'some_key_1':np.array([1,2,3,4]), 'some_key_2':np.array([2,3,4,5])}
print(np.array(list(data.values())))
# [[1 2 3 4]
# [2 3 4 5]]
You get a 2-D array, with rows in a random order (Python dicts are unordered).
To get a 1-D array, you can use np.concatenate:
print(np.concatenate(list(data.values())))
# [1 2 3 4 2 3 4 5]
If the order is important, you could use a list of tuple or an OrderedDict instead of a dict.
Probably the easiest way:
result = np.array(list(dict.values()))
With list, the dictionary's values will be stored in a list (i.e. one obtains a list where each item is a numpy array). This list of numpy arrays can then itself be converted to a numpy array by applying np.array().
You should not use dict as a name for your dictionary, though, as this shadows the built-in name dict. Instead, use e.g.:
result = np.array(list(dct.values()))
And make sure to rename your dictionary to dct.
You can get all values of a dictionary by using dictionary.values() and pass it to numpy.array
import numpy as np
np.array(list(dict.values()))
Please do not use dict as name it is a reserved keyword in python
Example
import numpy as np
data = {'some_key_1':np.array([1,2,3,4]), 'some_key_2':np.array([2,3,4,5])}
np.array(list(data.values()))
Output
array([[2, 3, 4, 5],
[1, 2, 3, 4]])
np.array(dict.values())? If you have a ragged number of elems that could be an issue.vstack, have a look :)