What is the right way to store random numbers in my database?

Question

I am working on an application which will generate unique random numbers and then store them into a database. I will check if a number exists through a HTTP request. Initially, for getting started, I would use around 10,000 numbers.

Is this the right approach?

Generate a random number, and, one by one, store them into an array and continue checking for array uniqueness, and when the array is complete, store the whole array to the database after sorting it.
Use the database and check to see if a number exists or not.

Which database should I use, as the application can scale up to 1 million numbers.

score 1 · Accepted Answer · edited May 23 '17 at 12:12

1

It may be more efficient, particularly if you want to generate 1000000 numbers, to make them one at a time and use validations in the model/database prevent duplicates.

As regards choosing a database, it will depend a little on you intended application. There is some info here: Which is the Best database for Rails application?

I can't comment on using a database directly from ruby without rails because I have not done that. One of the big pluses for rails for me is how easy it makes creating apps that use a database.

edited May 23 '17 at 12:12

Community

1
1

answered Apr 27 '13 at 09:39

Ian Kenney

6,376
1
25
44

So then how would i sort the database .. to make search fast. – Deepender Singla Apr 27 '13 at 10:10
Add an index to the column that you want to sort on – Ian Kenney Apr 27 '13 at 10:16

score 1 · Answer 2 · edited May 23 '17 at 12:12

A couple thoughts:

If you are storing 10 or 10,000 "random" numbers, what difference does it make whether they are random going into the database, or if the database randomly picks one number of a range of 10,000 sequential numbers? Do you need doubly-random number selections? MySQL, PostgreSQL and other DBMs can generate random numbers, and you can use their random number generator to retrieve a row, so you could either have it return a value directly from its generator, or grab a row. Either way, you don't need to worry about Ruby creating a random value -- unless you really want "triplely"-random numbers. I'd just stick the values of a (1..10_000) range into the database and call that part done and work on a query to grab records randomly.
If you want truly random numbers, you can't guarantee uniqueness. If you're happy with pseudo-random, you still have a problem because you could end up returning duplicates from inside the range unless you track which numbers you've used previously for a particular session. How you track uniqueness across a bunch of sessions is going to be an interesting problem if your site gets popular.

If I was doing this, I'd reverse some of the process. I wouldn't store the "random" values in the database, I'd use Ruby's built-in random number generator, and then probably check the database to see if I'd previously generated that number for that particular session. Overall, fewer values would be stored in the database so lookups to determine uniqueness would happen faster.

That would still be an awkward system to code and would grow inefficient over time as the "unique" records for sessions grew.

To do this without a database I'd create the random/unique range using something like: array = (1..10_000).to_a.shuffle, then each time I needed a value I'd use pop to pull the last value from the randomized array. I'd be tempted to pull from that pool of values for all sessions until it was exhausted, then regenerate it. There'd be a possibility of duplicate "unique" values at that point, but there should be a pretty small chance of the same number reappearing twice in a row.

What is the right way to store random numbers in my database?

2 Answers2