[crackmonkey] [scheide@usfca.edu: Adding one to a java hashtable]

[Seth David Schoen]

Nick Moffitt writes:

> ----- Forwarded message from Eric Scheide <scheide at usfca.edu> -----
> 
> Ok start thinking java.
> 
> words is the hash
> w is the key (a word)
> Look at the stupid crap you have to do to increment the Integer object.
> There must be an easier way.  Far as I know a hash table has to take an
> object as the value so you cant just use an int number.  It must be an
> Interger object.  Seems silly.
> 
>  if (words.containsKey(w)) {
>    words.put(w,new Integer(((Integer)words.get(w)).intValue()+ 1));
>  }

Well, you're not really using OOP that well to abstract out the nature of
this operation.

If you can store any kind of object in a hash, an assumption that seems
reasonable but whose correctness I'm really not positive of, you could do

public class incrementableInPlaceInteger{

int number;

public incrementableInPlaceInteger(){
	// Constructor
	number = 0;
}

public int value(){
	return number;
}

public void increment(){
	number += 1;
}

}

OK, now you store a "new incrementableInPlaceInteger()" in the hash.

When you want to increment one, you do

if (words.containsKey(w)) {
    ((incrementableInPlaceInteger)words.get(w)).increment();
  }

which is fully 8 bytes shorter than your version.  Or you can make a new
"incrementableInPlaceIntegerHash" descended from hash (which is then a hash
whose get method has a built-in typecast to force the value returned by the
get method into an incrementableInPlaceInteger).  Once you've written that
class, you can omit the typecast when actually making use of one:

if (words.containsKey(w)) {
    words.get(w).increment();
  }

Another option would be to write a hash class with a conditional get, or a
getOrInsert, or even a big method to do this whole thing:

words.getAndIncrementOrInsertWithIntegerValueInitializatedToZero(w);

-- by which time you've rewritten your Perl hash. :-)

But hey, you'd have written a pretty nice long method name...

(The parsing of "pretty nice long method name" is intentionally ambiguous.)

-- 
   Seth David Schoen L&S '01 (undeclared) / schoen at uclink4.berkeley.edu
He said, "This is what the king who will reign over you will do."  And they
said, "Nay, but we will have a king over us, that we also may be like all the
nations." (1 Sam 8)  http://ishmael.geecs.org/~sigma/   http://www.loyalty.org/
=====================================================================
To unsubscribe, send a message to majordomo at zork.net saying
"unsubscribe crackmonkey" in the body of the message.