I have an array with three kinds of value:
a = np.array([150, 50, 200, 50, 150])
How do I replace the values with ordered integers so that 50's are replaced with 0's, 150's with 1's, and 200's with 2's like so:
[1, 0, 2, 0, 1]?
Thank you.
You can just use integer division for your requirement.
a = a // 100
>>> a
array([1, 0, 2, 0, 1], dtype=int32)
You can use np.vectorize if you have to map different numbers to different keys
Ex:
a = np.array([150, 50, 200, 50, 150, 300])
check_value = {50:0, 150: 1, 200:2}
print(np.vectorize(lambda x: check_value.get(x, x))(a))
Output:
[ 1 0 2 0 1 300]
Assuming that your intention is to replace elements with their corresponding rank when sorted..
# sample data
a = np.array([150, 50, 200, 50, 150])
# get the rank of the elment
rank_list = np.unique(np.sort(a))
# get the index of the element in the rank list
output = [np.where(rank_list==x)[0][0] for x in a]
output will give you [1, 0, 2, 0, 1]
np.floor(a / 100).astype(int)