Get File Extension for special cases like tar.gz

Question

I need to extract extensions from file names.

I know this can be done for single extensions like .gz or .tar by using filePath.lastIndexOf('.') or using utility methods like FilenameUtils.getExtension(filePath) from Apache commons-io.

But, what if I have a file with an extension like .tar.gz? How can I manage files with extensions that contain . characters?

Actually, the extension is not `.tar.gz`. The extension - by definition - is `.gz`. The `.tar` part is only for informational purposes. If the file ended with `.test.zip` you would consider the extension `.zip`, wouldn't you? — Thorsten Dittmar, Jul 16 '13 at 12:38
and `commons-io-2.4.jar` in this case what you would expect ? — vels4j, Jul 16 '13 at 12:40
@ThorstenDittmar It is still reasonable to handle *.tar.gz reasonable, just like *.tgz, mind you. — Ingo, Jul 16 '13 at 12:40
oh it's true I haven't realized that @ThorstenDittmar. So .tar.gz and .gz both have the same File properties (e.g. icon) in this case? — Bernice, Jul 16 '13 at 12:40
If you don't want to use commons-io, there is a similar method in Guava; `Files.getFileExtension(filePath)` - but it uses the same `lastIndexOf('.')` technique. — grkvlt, Jul 21 '13 at 14:09
Why has this been closed? The question is reasonable, and *demonstrate[s] a minimal understanding of the problem* as far as I can see. Certainly, there was enough information for me to provide an answer, which was then accepted. I have edited the question slightly to clarify this, and the edit has been accepted. — grkvlt, Jul 22 '13 at 13:29
@vels4j In that case `commons-io-2.4.jar` has the extension `jar`. For other extensions, the mappings should look like this: `commons-io-2.4.tar` = `tar`, `commons-io-2.4.tar.bz2` = `tar.bz2`, `commons-io-2.4.tgz.shar` = `tgz.shar`, `commons-io-2.4.tgz` = `tgz`, `commons-io-2.4.jar.md5` = `md5`. Note the last one, this is an MD5 signature of the Jar file, so the extension is `md5` not `jar.md5`. — grkvlt, Jul 22 '13 at 13:35
@grkvlt I knew, but what if you get an unknown type. That is how this question is closed. — vels4j, Jul 22 '13 at 14:04

score 5 · Accepted Answer · answered Jul 16 '13 at 13:58

If you know what extensions are important, you can simply check for them explicitly. You would have a collection of known extensions, like this:

List<String> EXTS = Arrays.asList("tar.gz", "tgz", "gz", "zip");

You could get the (first) longest matching extension like this:

String getExtension(String fileName) {
  String found = null;
  for (String ext : EXTS) {
    if (fileName.endsWith("." + ext)) {
      if (found == null || found.length() < ext.length()) {
        found = ext;
      }
    }
  }
  return found;
}

So calling getExtension("file.tar.gz") would return "tar.gz".

If you have mixed-case names, perhaps try changing the check to filename.toLowerCase().endsWith("." + ext) inside the loop.

score 4 · Answer 2 · answered Jul 16 '13 at 12:49

4

A file can just have one extension!

If you have a file test.tar.gz,

.gz is the extension and
test.tar is the Basename!

.tar in this case is part of the basename, not the part of the extension!

If you like to have a file encoded as tar and gz you should call it .tgz. To use a .tar.gz is bad practice, if you need to handle thesse files you should make a workaround like rename the file to test.tgz.

answered Jul 16 '13 at 12:49

Grim

1,938
10
56
123

1

Beg to differ: the "extension" is whatever some application defines it to be. From a OS point of view there is no such thing as an "extension". – Ingo Jul 16 '13 at 13:41
A OS pov, hm ... its maybe more save to say from FS pov. – Grim Jul 16 '13 at 13:54
1

this is not helping if you are renaming your file to uuid + extension, you will get from `foo.tar.gz` to `61822326-ef4d-49f4-971d-b20269c72db9.gz` ... – Enerccio Jul 19 '18 at 17:52

score 0 · Answer 3 · answered Jul 16 '13 at 12:36

0

Found a simple way. Use substring to get filename only and indexOf instead of lastIndexOf to get first '.' and extension after it

answered Jul 16 '13 at 12:36

Bernice

2,552
11
42
74

2

So what would your program consider the extension if I gave you a file `program.logs.tar.gz`? – Thorsten Dittmar Jul 16 '13 at 12:40
Agree, this won't work for files with things like version numbers embedded in them. – grkvlt Jul 16 '13 at 14:07

score 0 · Answer 4 · answered Jul 16 '13 at 12:37

You can get the filename part of the path, split on . and take the final 0, 1, or 2 elements in the array as the extension.

Of course if .tar.* (gz, bz2, etc.) is your only edge case it may be pragmatic to just build a solution that filters filenames for .tar. and use that as the point at which to extract the extension (to include the .tar portion).

Get File Extension for special cases like tar.gz

4 Answers4