how is the "find" algorithm in text editors constructed?

[zskillz]

I'm going to have to write a program that parses through a very large text file looking for an instance of a specific word....

in the past, I've just read it line by line, and searched each token for a match to what I want... however, this seems like it'd be really really slow. So I'm curious how programs like wordpad/notepad, etc. do it. I feel like I could really cut down on a lot of wasted time if I could mimic that implementation...

so I'm looking for suggestions or any sort of help that you have to offer!

thanks in advance,
-Z

[SpookyEddy]

I would select your open source text editor of choice, grab its source code and have a quick look how they implement search functions. I had a quick check and vim, emacs and nano all have a search.c that should give you some ideas, link lists seem to be mentioned.

Regards

ed

[pedantic]

however, this seems like it'd be really really slow.

was it really slow in the past? like, when you did it?

text editors simply search the string/stream/char_array that is in memory. no reason for file/io, just memory search.

[Nighthawk]

It depends on the language you're using. I think the way I'd do it is to write it in Perl using a regular expression and make it into a module.. I'm not sure whether other languages have regexp or not.

[zskillz]

So I'm doing this in java right now, but I'll probably convert it to C or C++ (more likely) in the future.

The files that I'm going to be searching through are going to be multiple thousands of lines of code... possibly up to 100k+ lines. (these are the output files of electronic structure calculations on various molecules). The output is so long because this technique is a self-consistant field (SCF) calculation, so it takes the results from the current run - prints them - then reinserts them into the new run. These repeats until the convergence tolerence is reached....

anyway... it's sort of beside the point. I'm looking for a way to rip a chunk of the output from the code (the final output actual) and then I'll be doing calculations based on the output that I take....

I was just hoping to find a faster/more efficient way to get to the information that I want then going line by line through these gigantic files.

any other ideas are invited and appreciated.

thanks
-Z

[pedantic]

any other ideas are invited and appreciated.

except mine...

I was just hoping to find a faster/more efficient way to get to the information that I want then going line by line through these gigantic files.

stop searching, line by line, read the gigantic file at once, then search for token!

how slow was it in the past?

[VHockey86]

Try searching the web for source code on "find" functions