glob (programming)

From Wikipedia, the free encyclopedia
Jump to: navigation, search

In computer programming, in particular in a Unix-like environment, the term globbing is sometimes used to refer to pattern matching based on wildcard characters.[citation needed] The noun "glob" is used to refer to a particular pattern, e.g. "use the glob *.log to match all those log files".[citation needed] Its notation is simpler than regular expressions, and without their expressive power.

Origin[edit]

The command interpreters of the early versions of Unix (1st through 6th Editions, 1969–75) did not expand wildcard characters in file path arguments to a command; a separate program, /etc/glob,[1] performed the expansion and supplied the expanded list of file paths to the command for execution. Its name is an abbreviation for "global command".[2] Later, this functionality was provided as a library function, glob(), used by programs such as the shell.

Technical[edit]

Unix shell globbing operates by parameter expansion – the glob pattern (e.g., *.log) is expanded and replaced by the list of all matches. For example, if a directory contains two files, a.log and b.log then the command cat *.log will be expanded by the shell to cat a.log b.log which is then evaluated (in this case, displaying the files). The order of arguments to a command often matters – for example, cat a.log b.log prints first a.log and then b.log, while cat b.log a.log prints first b.log and then a.log. Thus, while "filenames that match the pattern" is an (unordered) set, the actual expanded list of matching files is an ordered list, a sequence, and thus an order must be chosen, conventionally alphabetical order, however defined by the shell.[3]

Implementations[edit]

Unix shells such as Bash, tcsh, and zsh provide globbing on filenames at the command line and in shell scripts.[4]

The Windows command interpreter cmd.exe relies on a runtime function in applications to perform globbing.[5][6] Windows PowerShell Cmdlets support globbing.[7]

The term "glob" is also used to refer more generally to limited pattern-matching facilities of this kind, in other contexts:

  • Go has a Glob function in the filepath package.[8]
  • Java has a Files class containing methods that operate on glob patterns.[9]
  • Haskell has a Glob package with the main module System.FilePath.Glob. The pattern syntax is based on a subset of Zsh’s. It tries to optimize the given pattern and should be noticeably faster than a naïve character-by-character matcher.[10]
  • Perl has both a glob function (as discussed in Larry Wall's book Programming Perl) and a Glob extension which mimics the BSD glob routine.[11] Perl's angle brackets can be used to glob as well: <*.log>.
  • PHP has a glob function.[12]
  • Python has a glob module in the standard library which performs wildcard pattern matching on filenames.[13] and an fnmatch module with functions or matching strings or filtering lists based on these same wildcard patterns [14] Guido van Rossum, author of the Python programming language wrote and contributed a glob() routine to BSD Unix in 1986.[15] There were previous implementations of glob, e.g., in the ex and ftp programs in previous releases of BSD.
  • Ruby has a glob method for the Dir class which performs wildcard pattern matching on filenames.[16] Several libraries such as Rant and Rake provide a FileList class which has a glob method or use the method FileList.[] identically.
  • Tcl contains both true regular expression matching facilities and a more limited kind of pattern matching often described as globbing.[17]

Syntax[edit]

Although there is no definite syntax for globs, common features include:

Task Type Example Unix shells COMMAND.COM 4DOS, 4OS2, 4NT, TCC CMD.EXE Windows PowerShell SQL (LIKE operator) SAP
Match one or zero unknown characters Standard globbing, standard wildcards ?at matches at, Cat, cat, Bat or bat N/A ? ? ? N/A N/A N/A
Match exactly one unknown character Standard globbing, extended wildcards ?at matches Cat, cat, Bat or bat, but not at ? N/A [?] N/A ? _ +
Match no character (match empty set) Standard globbing, extended wildcards Letter[] matches Letter, but not Letter1 Unknown N/A [] or [!?] N/A Unknown Unknown Unknown
Match any number of unknown characters from the position in which it appears to the end of the filename or extension Standard globbing, standard wildcards Law* matches Law, Laws, or Lawyer * * * * * % *
Match any number of unknown characters (regardless of the position where it appears, including at the start and/or multiple times) Standard globbing, extended wildcards *Law* matches Law, GrokLaw, or Lawyer * N/A * N/A * % *
Match a character as part of a group of characters Standard globbing, extended wildcards [CB]at matches Cat or Bat but not cat or bat [characters] N/A [characters] N/A [characters] N/A[18] Unknown
Match any character but the one specified Standard globbing, extended wildcards [!C]at matches Bat, bat, or cat but not Cat [!characters] N/A [!characters] N/A N/A N/A Unknown
Match a character as part of a character range Standard globbing, extended wildcards Letter[0-9] matches Letter0, Letter1 etc. but not Letters or Letter Unknown N/A [a-b] N/A Unknown Unknown Unknown
Match any character but the range specified Standard globbing, extended wildcards Letter[!3-5] matches Letter1, Letter2, but not Letter3, Letter4 or Letter5 Unknown N/A [!a-b] N/A Unknown Unknown Unknown
Escape character Standard globbing, extended wildcards Law\* will only match Law* \ N/A %= internal variable, configurable by SETDOS /E command or EscapeChar= config directive (default for 4DOS: (0x18), else: ^) ^ ` N/A N/A

Some shells (like the C shell) support additional syntax including alternation or brace expansion, also known as extended globbing.

Globs do not include syntax for the Kleene star which allows multiple repetitions of the preceding part of the expression; thus they are not considered regular expressions, which can describe the full set of regular languages over any given finite alphabet.[citation needed]

Standard SQL uses a glob-like syntax for simple string matching in its LIKE operator. The percent sign (%) matches zero or more characters, and the underscore matches exactly one character. The term "glob" is not generally used in the SQL community, however. Many implementations of SQL have extended the LIKE operator to allow a richer pattern-matching language incorporating elements of regular expressions.

See also[edit]

References[edit]

  1. ^ "First Edition Unix manual 'Miscellaneous' section (PDF)" (PDF). Retrieved 2011-05-11. 
  2. ^ 1st Edition UNIX, code.google.com, src/cmd/glob.c 
  3. ^ "The Open Group Base Specifications Issue 6, IEEE Std 1003.1, 2004 Edition, 2.13.3 'Patterns Used for Filename Expansion'". 
  4. ^ The "Advanced Bash-Scripting Guide, Chapter 19.2: Globbing" (Mendel Cooper, 2003) has a concise set of examples of filename globbing patterns.
  5. ^ "Wildcard Expansion". Microsoft Developer Network. 2013. 
  6. ^ "Expanding Wildcard Arguments". Microsoft Developer Network. 2013. 
  7. ^ "Supporting Wildcard Characters in Cmdlet Parameters". Microsoft Developer Network. 2013. 
  8. ^ "Package filepath - The Go Programming Language". Golang.org. Retrieved 2011-05-11. 
  9. ^ "File Operations". Oracle. Retrieved 2013-12-16. 
  10. ^ "Glob-0.7.4: Globbing library". Retrieved 2014-05-07. 
  11. ^ Contact details. "File::Glob - Perl extension for BSD glob routine". perldoc.perl.org. Retrieved 2011-05-11. 
  12. ^ "glob - Manual". PHP. 2011-05-06. Retrieved 2011-05-11. 
  13. ^ "10.7. glob — Unix style pathname pattern expansion — Python v2.7.1 documentation". Docs.python.org. Retrieved 2011-05-11. 
  14. ^ "10.8 fnmatch Unix filename pattern matching -- Python v2.7.7 documentation". Docs.python.org. Retrieved 2014-06-28. 
  15. ^ "'Globbing' library routine". Archived from the original on 2007-12-19. Retrieved 2011-05-11. 
  16. ^ "Class: Dir". Ruby-doc.org. Retrieved 2011-05-11. 
  17. ^ "TCL glob manual page". Retrieved 16 November 2011. 
  18. ^ Some proprietary extensions such as Transact-SQL provide this functionality, e.g., [characters] in as discussed in LIKE (Transact-SQL)