Generate Random Integers

Is there a way that column A and B has 1 duplicate numbers only? like in this sample below values i put 1 to 10

# A B
1 3 10
2 5 6
3 3 8
4 2 4
5 1 9
6 7 9
7 5 6
8 2 10
9 1 8
10 4 7

I am not sure what you mean by "1 duplicate only". Assuming that you want a list of random numbers where each value appears no more than twice, I tried this "iterative" technique:

1) Generate 50 random numbers between 1 and 70. In A1, enter =RANDBETWEEN(1,70) -- copy down to A50
2) Count duplicates in column B. B1 =COUNTIFS(A$1:A1,A1). Copy down to B50.
3) Generate an alternate random number for those that are duplicated too many times. C1 =RANDBETWEEN(1,70) [a copy of A1].
4) Combine list into new list. D1 =IF(B1>2,C1,A1)
5) Check for duplicates in new list. E1 =COUNTIFS(D$1:D1,D1) [a copy of B1].
6) Generate another alternate random number in column F. F1 =RANDBETWEEN(1,70) [a copy of A1].
7) In column G, combine lists G1 =IF(E1>2,F1,D1) [a copy of D1].

Testing it in my spreadsheet, 3 "iterations" never had a value appear more than twice. The random nature of the algorithm suggests that the probability of the third iteration having a value appear more than twice is not 0, so one would need to test this enough to know how many iterations are needed to be confident enough that it will not generate a list of random numbers where one number appears more than twice. If you pay attention to relative and absolute references, additional iterations can be achieved by continuing to copy the groups of 3 columns across.

Did I understand correctly?

Thank you for immediate response, I do learned a lot from you regarding iterative technique.

This is my project i working on. I don't know if its called generate "random" sequence of numbers or generate sequence of numbers.

In this sample of below would like to seek formula for as follows;

*=C2:F24 are the range of cells that create sequence of numbers
*Column C to F must have order number (smaller to higher), C is smaller until F is higher number randomly.
*List of number appear no more than twice.
*Column G to I is difference of between the number on column C to F, That will generate a random pairing of numbers in (C2,D2 and E2 whose sum is in G2,H2 and I2).

TestSequencing.xls
TestSequencing.xlsx

I'm still not sure I understand. Looking at your latest sample, I wonder if it is necessary to generate the random numbers in C2:F2? My first thought is that it might be easier to generate the differences (G:I) randomly, then compute the sequence (C:F) from the random differences:

1) Generate a random starting number in C2 =RANDBETWEEN(1:15) I am guessing at the lower and upper limits here.
2) Generate a random difference in G2 =RANDBETWEEN(5,20). Again, I am just guessing what the lower and upper limits ought to be. Copy G2 across to I2.
3) D2 becomes =C2+G2, copy across to F2
4) Select C2:I2 and copy down to row 24.

I don't see a ready way to prevent a sequence from occurring more than twice, so one would need to count each sequence of numbers, test if any of the sequences occur more than twice, then replace any such sequences with a new sequence. Test again, until all sequences occur no more than twice.

Is that an allowed algorithm? Will it have the correct properties and distribution?

Yes, that was my type. It should be RANDBETWEEN(1,15).