in reply to Random hash element (low memory edition).

You can find the number of elements using keys in a scalar context, and it won't take time or space to go through the elements.

But you could use the recipe for printing a random line from a file to just iterate through the hash without knowing how many elements it has and select a random one.

Yes, as long as you don't change the hash, keys will return keys in the same order every time. values will return values in the corresponding order.

Caution: Contents may have been coded under pressure.

Replies are listed 'Best First'.
Re^2: Random hash element (low memory edition).
by amarquis (Curate) on Jan 24, 2008 at 20:51 UTC

    That method is awfully neat, but my perl skills are not so I'll hit you with a follow-up question.

    rand($.) < 1 && ($line = $_) while <>;

    So this iterates through the file and, if a random number from 0 to the current line number is less than 1, selects the current line, overwriting previous choices? So, in my case, I'd just keep a count variable for how many elements I've been through and use that in place of $.?

    I'm going to have to check out the proof of that, it is really interesting that everything evens out (i.e. the first elements have a high chance of being overwritten, the latter elements have a high chance of just not being picked, and it all balances out to a random selection.) Thanks for pointing me to that FAQ entry!

[reply]
[d/l]
      So, in my case, I'd just keep a count variable for how many elements I've been through and use that in place of $.?

      Exactly.

      Caution: Contents may have been coded under pressure.
[reply]

In Section Seekers of Perl Wisdom