Bash Extended Globbing

 in

Wildcards in bash are referred to as pathname expansion. Pathname expansion is also sometimes referred to as globbing. Pathname expansion "expands" the "*", "?", and "[...]" syntaxes when you type them as part of a command, for example:

  $ ls *.jpg         # List all JPEG files
  $ ls ?.jpg         # List JPEG files with 1 char names (eg a.jpg, 1.jpg)
  $ rm [A-Z]*.jpg    # Remove JPEG files that start with a capital letter
A subtle point about pathname expansion that is not often understood is that it is done by bash and not by the operating system or by the program that is being run. The program never sees the wildcards, bash substitutes the expansion into the command line before running the program. This is rarely important except when you're writing code and calling exec() and friends: if you don't execute the program via bash any wildcards in the command line that you pass to exec() won't get expanded.

But these are not the only forms of wildcards supported by bash. The other forms are referred to as extended globbing and you must enable them before you can use them:

  $ shopt -s extglob
Extended globbing as described by the bash man page:
  ?(pattern-list)   Matches zero or one occurrence of the given patterns
  *(pattern-list)   Matches zero or more occurrences of the given patterns
  +(pattern-list)   Matches one or more occurrences of the given patterns
  @(pattern-list)   Matches one of the given patterns
  !(pattern-list)   Matches anything except one of the given patterns
Here a pattern-list is a list of items separated by a vertical bar "|" (aka the pipe symbol). If you look at these you can see why the leading character was chosen as it matches what is used in regular expression syntax:
  Bash              Regular Expression
  ?(pattern-list)   (...|...)?
  *(pattern-list)   (...|...)*
  +(pattern-list)   (...|...)+
  @(pattern-list)   (...|...)    [@ not a RE syntax]
  !(pattern-list)   "!" used as for negative assertions in RE syntax
Well, except for the "@" character you can see why...

For example, to list all the JPEG and GIF files that start with either "ab" or "def" you could do:

  $ ls +(ab|def)*+(.jpg|.gif)
Of course you could also do this without extended globbing:
  # ls ab*.jpg ab*.gif def*.jpg def*.gif

To list all the files that match the regular expression "ab(2|3)+.jpg" you could do:

  $ ls ab+(2|3).jpg
Now that's something you can't do with regular globbing. Note: this matches files like ab2.jpg, ab3.jpg, ab2222.jpg, ab333.jpg, etc.

However, probably the most interesting extended globbing syntax is the "!(...)" syntax for matching everything except (ie not matching). But, be careful, this doesn't always do what you might expect. For example, let's list all the files that aren't JPEGs or GIFs. One's first thought might be something like this:

  $ ls *!(.jpg|.gif)         # wrong, Wrong, WRONG
But that doesn't work because the ".jpg" and the ".gif" of any file's name end up getting matched by the "*" and the null string at the end of the file name is the part that ends up not matching the "!(...)" pattern. The correct syntax is:
  $ ls !(*.jpg|*.gif)

For a more complex negation example, let's go back to the first example and list all the files that aren't JPEG or GIF files and start with either "ab" or "def". This is actually quite simple, we just take the first example and nest it inside of "!(...)":

  $ ls !(+(ab|def)*+(.jpg|.gif))
Of course, like complex regular expressions, this will be completely incomprehensible 10 minutes after you write it.

______________________

Mitch Frazier is an Associate Editor for Linux Journal.

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Dangerous example for rm

Kevin Wilson's picture

In bash,

$ rm [A-Z]*.jpg # Remove JPEG files that start with a capital letter

removes [a-y]*.jpg files too if LANG is set to en_US! The z*.jpg files are ignored, because the collation order is AaBb..Zz. Bash documents that they'll use the collation order. Personally, I think they need to recognize that sorting for people and sorting for computers is different, and that upper case and lower ranges should always be separate for globbing. Admins are coding for computers, not people.

Not quite

Anonymous's picture

Actually, on my machine the order seems to be aAbB...zZ.

#~/mytest> ls [a-z]*
a A b B z
#~/mytest> ls [A-Z]*
A b B z Z

rm command's dangers

Pradeep's picture

A similar discussion is found here. Here, filenames having space in them, are not handled properly.

Same Results

teguh iskanto's picture

I could get the same result with :
(which I think it's much convenient / much easier to write)

- ls -alt [a-z]*.{jpg,gif}
- ls -alt [a-z]?.{jpg,gif}
- ls -alt t{[0-9]*,[a-z]*}.{jpg,gif}

hope this helps

Not really

Mitch Frazier's picture

This:

  ls -alt [a-z]*.{jpg,gif}

lists jpg/gif files that start with any lower case letter followed by any number of other characters.

This:

  ls -alt [a-z]?.{jpg,gif}

lists jpg/gif files start with any lower case letter followed by any single character.

This:

  ls -alt t{[0-9]*,[a-z]*}.{jpg,gif}

lists jpg/gif files that start with the letter t followed by either a digit or a lower case letter and then followed by any number of other characters.

Although the first two will list files that start with "ab" or "def", they also list many other files. Not even sure what your last example is meant to compare to since I don't have any examples that list files starting with the letter t.

The point of the examples is not to show the "simplest" way to perform the examples themselves, but rather to illustrate extended globbing.

Mitch Frazier is an Associate Editor for Linux Journal.

similar effect with another trick

Anonymous's picture

That:
$ ls +(ab|def)*+(.jpg|.gif)
can be done in bash NOT with:
$ ls ab*.jpg ab*.gif def*.jpg def*.gif
but more effieciently with:
$ ls {ab,def}*.{jpg,gif}

Though you missed the point. 1st syntax is different from 2nd (and 3rd) in that in 2nd and 3rd cases "ls" would complain if some of the patterns are not found. (e.g. "ab*.jpg" not found) while in first case bash would properly expand the expression listing only matching files.

Only if you squint

Mitch Frazier's picture

True, they do behave differently if there are no matching files. And yes the "syntax" is different since they are written differently, what you probably mean is that the "semantics" are different because they do not act the same if the expansion fails. So, when I stated Of course you could also do this without extended globbing, it was not my intention to imply that the two forms were exactly the same in all instances. The point here was to illustrate how to use extended globbing and not to provide a comparison between the different ways to list jpg/gif files that start with "ab" or "def".

Mitch Frazier is an Associate Editor for Linux Journal.

White Paper
Linux Management with Red Hat Satellite: Measuring Business Impact and ROI

Linux has become a key foundation for supporting today's rapidly growing IT environments. Linux is being used to deploy business applications and databases, trading on its reputation as a low-cost operating environment. For many IT organizations, Linux is a mainstay for deploying Web servers and has evolved from handling basic file, print, and utility workloads to running mission-critical applications and databases, physically, virtually, and in the cloud. As Linux grows in importance in terms of value to the business, managing Linux environments to high standards of service quality — availability, security, and performance — becomes an essential requirement for business success.

Learn More

Sponsored by Red Hat

White Paper
Private PaaS for the Agile Enterprise

If you already use virtualized infrastructure, you are well on your way to leveraging the power of the cloud. Virtualization offers the promise of limitless resources, but how do you manage that scalability when your DevOps team doesn’t scale? In today’s hypercompetitive markets, fast results can make a difference between leading the pack vs. obsolescence. Organizations need more benefits from cloud computing than just raw resources. They need agility, flexibility, convenience, ROI, and control.

Stackato private Platform-as-a-Service technology from ActiveState extends your private cloud infrastructure by creating a private PaaS to provide on-demand availability, flexibility, control, and ultimately, faster time-to-market for your enterprise.

Learn More

Sponsored by ActiveState