I compute an array of indices in each iteration of a loop and then I want to remove the duplicate elements and concatenate the computed array to the previous one. For example the first iteration gives me this array:
array([ 1, 6, 56, 120, 162, 170, 176, 179, 197, 204])
and the second one:
array([ 29, 31, 56, 104, 162, 170, 176, 179, 197, 204])
and so on. How could I do it?
you can concatenate arrays first with numpy.concatenate then use np.unique
import numpy as np
a=np.array([1,6,56,120,162,170,176,179,197,204])
b=np.array([29,31,56,104,162,170,176,179,197,204])
new_array = np.unique(np.concatenate((a,b),0))
print new_array
result:
[ 1 6 29 31 56 104 120 162 170 176 179 197 204]
numpy.union1d is the one-numpy function you're looking for (in this case).
import numpy as np
print( np.union1d([ 1, 6,56,120,162,170,176,179,197,204],\
[29,31,56,104,162,170,176,179,197,204]) )
# result
[ 1 6 29 31 56 104 120 162 170 176 179 197 204]
You can use numpy.concatenate and numpy.unique:
In [81]: arr = np.array([ 1, 6, 56, 120, 162, 170, 176, 179, 197, 204])
In [82]: arr = np.unique(np.concatenate((arr, np.array([ 29, 31, 56, 104, 162, 170, 176, 179, 197, 204]))))
In [83]: arr
Out[83]: array([ 1, 6, 29, 31, 56, 104, 120, 162, 170, 176, 179, 197, 204])