Re: offtopic :)

From: Jason Wilkins (fenix@IO.COM)
Date: 10/03/98

Ronny Iversen wrote:

> Hi :)
> Can anybody tell me whats the most efficient way of reading a file,
> stripping a line from it, and writing  the new result over
> the old file. Is it possible to do this without having to read the
> entire file into an array into memory first?

Surprisingly, it the way to do this is to reduce the number of
reads and writes to the absolute minium.  If you need to search
the entire file then read the whole file in one go.  I/O operations
to disk take much longer than writing to memory (millions of
times longer!).  So, if the file is not larger than a a couple hundred
K your best bet is to do it in one read and one write.

So, the questions are, how often do you have to do this and how
big the file is.  If you only do it once in a while then by all means
just read the whole file into memory and then just write it back out.
If you do it more then read the file into memory, remove all the lines
need to, and then write it back periodically or only when you have
idle time.  This would reduce it to less than one read and write
per line removed.

If the file is really large then keeping it in memory, or reading it all

at once is a problem.  The only way to optimize it would be to
use a more structured format than a text file.  The file would have
to be divided into chunks that each can be loaded individually
into memory.

You see the hard drive has to load/save a whole sector no matter
how little the data is.  So it could take just as long to load a byte
as to load a 64K file.  If you have a small file load it all.  If its
restructure so that it can be divided into chunks that can be loaded
all at once.

> Please gimmme some oseudo code or something :)

I can't give you any psuedo code unless I know how often you are
doing this and how big the file is.

> Regardz

           Phoenix -- President of The Artistic Intuition Company
       Caelius * Mirror Reflex * Runica * X-Domain * Infinite Realms

     | Ensure that you have read the CircleMUD Mailing List FAQ:  |
     | |

This archive was generated by hypermail 2b30 : 12/15/00 PST