AutoHotkey Community

[toralf]

I want my script to treat binary files differently than ASCII files . How can I find out if a file contains binary or ASCII data?
_________________
Ciao
toralf

[BoBo]

D'ya wanna transfer (FTP) your files via an AHK script ?

[toralf]

No, i want to compress the files differently.
_________________
Ciao
toralf

[TeknoMusicMan]

try using FileRead to read the file, then check it against a IF statement to see if any of the characters are other then 0 or 1, if not its binary, if it does its ascii.

unless you mean like executables and such, in that case would the .XXX extension be enough to tell you what it is. You could setup a list of binary extensions and a list of ascii extensions and filter them that way.

[Titan]

^ example:

[toralf]

Thanks you all for your comments.

FileRead is not an option, since there are several hundered files with up to 3 GB (!) filesize. So a loop over all files will be too costly.

From all your post I assume the only way to do is to look directly at the contens. So I will try to read 5 or 10 of the first lines (do they exist in binary format?) and guess for the rest.
_________________
Ciao
toralf

[Titan]

Hmm, reading only a few random lines of a file would be a clever idea. Why couldn't you just take .bin files for binary?
_________________

[Chris]

If it weren't for the fact that some of your files are so large, the following FileRead idea would probably work. But since they are so large, I think it's impossible to do it directly inside a script. You would probably need an external utility that scans the file for the first binary zero and stops there, reporting the file as binary.

Method that won't help for huge files, but you could use it for ones that pass a smallness filter:
Due to the fact that binary zeroes are used as string terminators in AutoHotkey, if you use FileRead to read in the file and StringLen says the variable's contents are lot shorter than FileGetSize says they should be, that probably means the file is binary.

[toralf]

The files I need to process are created by another software that doesn't give extensions to them.
_________________
Ciao
toralf

[TeknoMusicMan]

Does this software have any sort of naming standard for the files?

what i mean is there any pattern in the filenames or the binary files and non-binary files?

IE: 04152005bin or anything of that nature. If so you could read all the filenames in a directory into an array and read each name to see if it fits the pattern for a bin or non-bin

[BoBo]

This might be of help: [file - Dateityp ermitteln] (German). Now you've to check if file is available as an Windows based ---> UnixTool, like [here]

[BoBo]

If Rabiator is right, as said in [this post] AHK's FileRead could be an option.

[toralf]

Thanks BoBo,
Chris suggested the same thing 4 posts above yours. :)
_________________
Ciao
toralf

[BoBo]

I've realized that Chris has mentioned that "workarround"

hm - while Chris pointed out to check for the vars length (StringLen), I've assumed Rabiator's hint is about FileRead's errorlevel. But - who cares

[toralf]

When I read the manual for FileRead I don't think the Errorlevel will do th etrick, because AHK "thinks" it had read the file successfully when it reaches the first binary 0.
But thanks for pointig that out, I'll look deeper into this soon.

Thank you all, very much appreaciated.
_________________
Ciao
toralf