Very Computer

damn man I think I'm a pretty decent programmer but I'd go 1 for 3 on these.
I guess I could stumble through a guess for the other two, but
it'd just be a guess based on how I'd do it as a programmer.
I'd probably just say 'I don't know' though to cut down on awkward time.

Herez my favorite interview question of all time:

I hand you a big file, I want to know how many times the string
'wakawaka' appears in it. How do you find that?

If they are right out of school, I might sit there and listen
to a description of an algorithm and be happy with it. Otw,
what I want to hear is something like:

more <file name> | grep wakawaka | wc

(ok the more isn't necessary if you are one of those wackos
who gets bent out of shape by this kind of usage).

Its so trivial for real unix geeks that you don't feel like you
are ambushing them, but its still amazingly good at separating
the wheat from the wannabes (to really badly mix my metaphors).
--
I used to think government was a necessary evil.
I'm not so sure about the necessary part anymore.

That won't find 'wakawaka' if it's broken over lines ...

And it if the line contains things other than wakawaka, wc will count
those as well in its summary of the number of words. Only useful thing
you'll get is the number of lines with a complete 'wakawaka' on them.

I'd have a crack at a perl script that used regular expressions,
personally. You could do it all one one line.

Let's ignore that case -- after all you do not know whether
"waka-\nwaka", "waka\nwaka" or "waka\n\rwaka" was intended to
be counted.

Quote:> And it if the line contains things other than wakawaka, wc will count
> those as well in its summary of the number of words. Only useful thing
> you'll get is the number of lines with a complete 'wakawaka' on them.

That is a bigger flaw in Chance's solution.

Quote:> I'd have a crack at a perl script that used regular expressions,
> personally. You could do it all one one line.

IMHO Perl would be an overkill:

    fmt -1 filename | grep wakawaka | wc -l

Cheers,
--
In order to understand recursion you must first understand recursion.

Nice.

I was thinking Perl to capture the wakawaka across broken lines. Isn't
it funny how there's always a utility *somewhere* in /usr/bin to do
exactly what you want ? :)

L8r

Luke.

But fmt  -1 won't separate two  occurance of wakawaka in  one word, as
in: wakawakawakawaka

What was asked is to count occurences of a string, not of a word.

So I would propose:

    cat file \

    | tr -d '\012' \
    | wc -c

--
__Pascal_Bourguignon__                   http://www.informatimago.com/
----------------------------------------------------------------------
 The name is Baud,...... James Baud.

Well, if we are so nitpicky, what is the correct answer to

  how many times does string 'wakawaka' occur in 'wakawakawakawaka' ?

I believe it is 3 (offsets 0, 4 and 8 ;-), not 2 as your command
reports.

I have no idea how to properly count those without awk, Perl or 'C' ...

Cheers,
--
In order to understand recursion you must first understand recursion.

You know, I just had a red hot go at doing it in perl and couldn't!! And
it's time to go home from work.

You could do (pseudocode)

x = 0
do
{
     match /wakawaka/ on line from offset x
     did we match a wakawaka already matched? (use a map here)
     no - increment match count

     increment offset x

Quote:} while (offset x < length of line);

But I wanted one regular expression that just did the lot!!

If you define wakawakawakawaka as two occurrences:

perl -lne'$c++while/wakawaka/g}{print$c'

If you define wakawakawakawaka as three occurrences:

perl -lne'$c++while$i=index($_,"wakawaka",$i)+1;$i=0}{print$c'

John
--
use Perl;
program
fulfillment

ya know, this is actually pretty slick. lots easier than trying
to figure out why my sed s/regexps/blabla/g never freakin work on the
first try.

BUT, I do feel obligated to point out that I never claimed
my solution was the answer. Just something LIKE what I would
want to hear if I got to ask this question in an interview.

Seems way better than "wheres the syntax error in this code?"
or "what flag do you pass to gnu make to do X", but still not
that hard to talk through to somebody with a clue.

--
I used to think government was a necessary evil.
I'm not so sure about the necessary part anymore.