From: Ron Jeffries <ronjeffries@...>
Date: 2002-03-17T10:58:27+09:00
Subject: sorting large string of records

Suppose I have a large string that is really a lot of fixed-size records joined
together. And suppose that I don't really want to turn it into an array of
strings.  (This last is a big supposition: I'm starting to think maybe I do want
to. But go with me ...)

I want to sort the records within the big string. 

Here's what I did because it was easy.

1. I have an object, StringView, that, given the string, an offset, and a
length, can act like a single record from the big string. I was already using it
to slice out fields from the "record". 

2. I implemented <=> on StringView, by slicing the string that it views and
sending <=> to that (and the string from the argument of course).

3. When the StringViews are sorted, I join('') them to get a new string
assembled.

This has the advantage of being easy but drawbacks include:

1. I have to slice the records out anyway
2. Slices occur more than N times, depending on the O() of Ruby's sort
algorithm.

Now if I were doing this in C or some language like that, I'd just use my
StringView as a pointer, and look at the bytes of the string without
instantiating a new object. A fairly simple loop over the characters could
implement <=>.

So ... is there a way to sort my big string without slicing it to smaller
strings?

Thanks!

Ronald E Jeffries
http://www.XProgramming.com
http://www.objectmentor.com
I'm giving the best advice I have. You get to decide whether it's true for you.