生成N个随机整数的数组，介于1和K之间，但每个数字至少包含一个

Question

I need to generate a matrix of N random integers between 1 and K , where each number appears at least once, having K ≤ N .

I have no problem using a call to numpy.random.random_integers() and checking the number of distinct elements, when K is much less than N , but it's harder to get a valid array when K approximates to N .

Is there any nice way to get this done? Any hint would be appreciated.

Answer 1

I don't think there's any built in functionality, but you could always make the exact array you want and shuffle/reshape.

Something like,

raw_array = np.append(np.arange(1, K+1), np.random.randint(1, K+1, size=(N-K)))
np.random.shuffle(raw_array)

This guarantees you have 1 to K from your arange, then the remaining required numbers (N - K) are filled by a random number from 1 to K.

Edited to specifically get 1 through K, rather than 0 to K - 1.

Answer 2

Fill the matrix iteratively with numbers 1-K so if K was 2, and N was 4, [1,2,1,2]. Then randomly generate 2 random numbers between 1-N where the numbers don't equal, and swap the numbers at those positions.

Answer 3

使用xrange（k）填充K个数字，然后使用随机数生成器填充（nk）个数字

Answer 4

您正在寻找np.random.choice(N, size=K, replace=False) doc ，它将返回K不同的数字，最多N 。