Beefy Boxes and Bandwidth Generously Provided by pair Networks
Clear questions and runnable code
get the best and fastest answer

Answer: How do I determine encoding format of a file ?

by donno20 (Sexton)
on May 09, 2003 at 01:53 UTC ( #256731=categorized answer: print w/replies, xml ) Need Help??

Q&A > files > How do I determine encoding format of a file ? - Answer contributed by donno20

Read the first two bytes of the file. Corresponding encoding and hex codes are as follow:
unicode Little Endian = "\xFF\xFE"
unicode Big Endian = "\xFE\xFF"
utf8 = "\xEF\xBB"
ASCII = straight to content

  • Comment on Answer: How do I determine encoding format of a file ?
Replies are listed 'Best First'.
Re: Answer: How do I determine encoding format of a file ?
by graff (Chancellor) on May 09, 2003 at 05:05 UTC
    Maybe in a perfect world... Certainly, there's a reasonable chance that UTF-16(BE|LE) really will start with the "byte-order-mark" (BOM, \x{FEFF}), but oh, so many people are not so reasonable. I've seen UTF16 files created by Microsoft tools that were little-endian (of course) but had no BOM.

    As for utf8 files, um, where did this information about "\xEF\xBB" come from? I've never seen a file that starts like that (and I would have thought that any proper utf8 mechanism would barf given this sort of byte sequence -- an initial "\xEF" would dictate the start of a 3-byte character, but you don't indicate a third byte). If you mean \x{EFBB} (expressed as UTF-16), this would be three octets when converted to utf8: \xEE\xBE\xBB).

    Regarding the notion of ASCII data, many people don't realize that ASCII files are simply a proper subset of utf8 files -- this was, I think, one of the design goals for utf8. (This adds to my doubts about "\xEF\xBB": this sequence isn't supposed to be in an ASCII file, yet an ASCII file is supposed to work as a utf8 file.)

    I think the question, though vaguely stated, may have been more concerned with distinguishing, say, the different flavors of ISO-8859 (which is impractical without knowledge of the language being used in the text, or at least some well-trained n-gram models for various languages), or CP12* vs. Mac* vs. 8859-* vs. euc-*, etc, etc (somewhat less speculative, but still not always simple or deterministic without modeling).

    Perhaps more clarification is needed about the scope of the question. ("raw" text files? HTML/XML files? files that are simply "some form of unicode, but I don't know which"?)

Log In?

What's my password?
Create A New User
and all is quiet...

How do I use this? | Other CB clients
Other Users?
Others perusing the Monastery: (4)
As of 2018-02-20 08:37 GMT
Find Nodes?
    Voting Booth?
    When it is dark outside I am happiest to see ...

    Results (268 votes). Check out past polls.