I've got the following code:

How can I reduce this to one file handle? I've got the developer's block ;-)

Note: What I'm trying to do is to get a number from a file, incrament the number by one, and rewrite the file data with the new number.

Cheers

Wil

- wil

I think you are using the best way really.

If you want one file handle you'd need +< but I don't find this very reliable

Could I ask why you have:

print ID2 ("$newkey");

Why not:

print ID2 $newkey;

?

I've written it as ("$newkey"); because I've taken that snippet out of a longer line of code, which wrote a number of lines to a flat file, and I wanted to have a "safer" way.

So, there is no way of doing this with one file handle? Hmm. There must be?

What was that +< you were talking about?

Thanks.

- wil

Yes theoretically you could use one filehandle using +< eg...

open F, "+</path/to/file" or die $!;

code

close F;

+< = read/write

Yeah, but I need the code :-)

- wil

It's the same as you have above except you only open/close once.

Hmm.

Doesn't seem to work.

Wil

- wil

Precisely...thats why I said


The way you have it is fine. It isn't going to slow your script down.

You'd make a good second-hand car salesman, Paul! ;-)

- wil

+< isn't going to help you, because when you open a file like that you generally shouldn't use <FH> to read from it - instead you need to use read(), which will read a certain number of bytes from the current position. Since it looks like you are simply trying to increment a counter on the first line, the line will be variable. If you really want to do it, you'll need to do something like this:


As you can see, it's simpler to just use two open()s.

---

Jason Rhinelander
Gossamer Threads
jason@gossamer-threads.com

Yeah, thanks Jason! I was just wondering on performance wise.

- wil

On a tiny file like that I think you'd be hard-pressed to find a noticeable difference. On a massive file, however, it becomes very useful.

---

Jason Rhinelander
Gossamer Threads
jason@gossamer-threads.com

So let me get this straight...

That code reads 4 bytes into the file and converts the bytes into a list of values taking the first one. Then we go back to the beginning of the file and the code adds one to the value we obtained and re-packs it into bytes and re-prints to the file?

Am I close?

Why 4 bytes?...is that the length of a character?

It's the length of a packed 32-bit integer. In this case, signed, but you could make it unsigned by change the "l" to "L" - still 4 bytes.

It's probably easier to think of the file as containing 32 bits, rather than 4 bytes. What is actually contained in the file is a binary number, which unpack converts into a decimal number, then it is incremented, then pack turns it back into binary and writes it to the file.

---

Jason Rhinelander
Gossamer Threads
jason@gossamer-threads.com

Ahhh. So I'm guessing it would be pretty tricky with a file with more data inside?

Say....

dsadas5dasdas

..or would you use a different method for that?

If your data consisted of strings, it would be pointless. But, if your data was all packed to certain lengths, this would generally be a faster (and easier) way to handle it. A text file or a file with variable length rows would become exceedingly difficult using this method, so you wouldn't use it.

---

Jason Rhinelander
Gossamer Threads
jason@gossamer-threads.com