Regular expressions, pattern matching 
Author Message
 Regular expressions, pattern matching

The python library documentation says that pattern matching is done like
Emacs, and as far as I can see, gives no further explanation. I don't have the
faintest idea how Emacs pattern matching works, or how the various pattern
matching modules/functions interact, so can someone point me to something like
the extensive Perl documentation.

Please don't point me to any Emacs doc. I don't use Emacs, and I don't want to
use Emacs, I just want to know how regular expressions and pattern matching
work in Python.
--
PJDM
Peter Mayne
Digital Equipment Corporation (Australia)
Canberra, ACT
----
These are my opinions, and have nothing to do with Digital.
The truth is out there, but not necessarily in here.



Thu, 14 May 1998 03:00:00 GMT  
 Regular expressions, pattern matching

Quote:

> Please don't point me to any Emacs doc. I don't use Emacs, and I don't
want to
> use Emacs, I just want to know how regular expressions and pattern matching
> work in Python.

Then you have a problem. Note that you don't need to use emacs in order
to use a part of its documentation. If you insist, you might try to use
the MOO doc on regexps that uses the same package (it's really the
same doc with the backslash escape changed to percent).

Just in case you change your mind, I append the essential part
of the emacs doc :-)

Regards,

Boris Borcic
==

Syntax of Regular Expressions
=============================

   Regular expressions have a syntax in which a few characters are
special constructs and the rest are "ordinary".  An ordinary character
is a simple regular expression which matches that character and nothing
else.  The special characters are `$', `^', `.', `*', `+', `?', `[',
`]' and `\'; no new special characters will be defined.  Any other
character appearing in a regular expression is ordinary, unless a `\'
precedes it.

   For example, `f' is not a special character, so it is ordinary, and
therefore `f' is a regular expression that matches the string `f' and
no other string.  (It does not match the string `ff'.)  Likewise, `o'
is a regular expression that matches only `o'.

   Any two regular expressions A and B can be concatenated.  The result
is a regular expression which matches a string if A matches some amount
of the beginning of that string and B matches the rest of the string.

   As a simple example, you can concatenate the regular expressions `f'
and `o' to get the regular expression `fo', which matches only the
string `fo'.  To do something nontrivial, you need to use one of the
following special characters:

`. (Period)'
     is a special character that matches any single character except a
     newline.  Using concatenation, you can make regular expressions
     like `a.b', which matches any three-character string which begins
     with `a' and ends with `b'.

`*'
     is not a construct by itself; it is a suffix, which means the
     preceding regular expression is to be repeated as many times as
     possible.  In `fo*', the `*' applies to the `o', so `fo*' matches
     one `f' followed by any number of `o's.  The case of zero `o's is
     allowed: `fo*' does match `f'.

     `*' always applies to the smallest possible preceding expression.
     Thus, `fo*' has a repeating `o', not a repeating `fo'.

     The matcher processes a `*' construct by immediately matching as
     many repetitions as it can find.  Then it continues with the rest
     of the pattern.  If that fails, backtracking occurs, discarding
     some of the matches of the `*'-modified construct in case that
     makes it possible to match the rest of the pattern.  For example,
     matching `ca*ar' against the string `caaar', the `a*' first tries
     to match all three `a's; but the rest of the pattern is `ar' and
     there is only `r' left to match, so this try fails.  The next
     alternative is for `a*' to match only two `a's.  With this choice,
     the rest of the regexp matches successfully.

`+'
     is a suffix character similar to `*' except that it requires that
     the preceding expression be matched at least once.  For example,
     `ca+r' will match the strings `car' and `caaaar' but not the
     string `cr', whereas `ca*r' would match all three strings.

`?'
     is a suffix character similar to `*' except that it can match the
     preceding expression either once or not at all.  For example,
     `ca?r' will match `car' or `cr'; nothing else.

`[ ... ]'
     `[' begins a "character set", which is terminated by a `]'.  In
     the simplest case, the characters between the two form the set.
     Thus, `[ad]' matches either one `a' or one `d', and `[ad]*'
     matches any string composed of just `a's and `d's (including the
     empty string), from which it follows that `c[ad]*r' matches `cr',
     `car', `cdr', `caddaar', etc.

     You can include character ranges in a character set by writing two
     characters with a `-' between them.  Thus, `[a-z]' matches any
     lower-case letter.  Ranges may be intermixed freely with
     individual characters, as in `[a-z$%.]', which matches any lower-
     case letter or `$', `%', or period.

     Note that inside a character set the usual special characters are
     not special any more.  A completely different set of special
     characters exists inside character sets: `]', `-', and `^'.

     To include a `]' in a character set, you must make it the first
     character.  For example, `[]a]' matches `]' or `a'.  To include a
     `-', write `---', which is a range containing only `-'.  To
     include `^', make it other than the first character in the set.

`[^ ... ]'
     `[^' begins a "complement character set", which matches any
     character except the ones specified.  Thus, `[^a-z0-9A-Z]' matches
     all characters except letters and digits.

     `^' is not special in a character set unless it is the first
     character.  The character following the `^' is treated as if it
     were first (`-' and `]' are not special there).

     Note that a complement character set can match a newline, unless
     newline is mentioned as one of the characters not to match.

`^'
     is a special character that matches the empty string, but only if
     at the beginning of a line in the text being matched.  Otherwise,
     it fails to match anything.  Thus, `^foo' matches a `foo' that
     occurs at the beginning of a line.

`$'
     is similar to `^' but matches only at the end of a line.  Thus,
     `xx*$' matches a string of one `x' or more at the end of a line.

`\'
     does two things: it quotes the special characters (including `\'),
     and it introduces additional special constructs.

     Because `\' quotes special characters, `\$' is a regular
     expression that matches only `$', and `\[' is a regular expression
     that matches only `[', and so on.

   Note: for historical compatibility, special characters are treated as
ordinary ones if they are in contexts where their special meanings make
no sense.  For example, `*foo' treats `*' as ordinary since there is no
preceding expression on which the `*' can act.  It is poor practice to
depend on this behavior; better to quote the special character anyway,
regardless of where is appears.

   Usually, `\' followed by any character matches only that character.
However, there are several exceptions: characters which, when preceded
by `\', are special constructs.  Such characters are always ordinary
when encountered on their own.  Here is a table of `\' constructs.

`\|'
     specifies an alternative.  Two regular expressions A and B with
     `\|' in between form an expression that matches anything A or B
     matches.

     Thus, `foo\|bar' matches either `foo' or `bar' but no other string.

     `\|' applies to the largest possible surrounding expressions.
     Only a surrounding `\( ... \)' grouping can limit the grouping
     power of `\|'.

     Full backtracking capability exists to handle multiple uses of
     `\|'.

`\( ... \)'
     is a grouping construct that serves three purposes:

       1. To enclose a set of `\|' alternatives for other operations.
          Thus, `\(foo\|bar\)x' matches either `foox' or `barx'.

       2. To enclose a complicated expression for the postfix `*' to
          operate on.  Thus, `ba\(na\)*' matches `bananana', etc., with
          any (zero or more) number of `na' strings.

       3. To mark a matched substring for future reference.

     This last application is not a consequence of the idea of a
     parenthetical grouping; it is a separate feature which happens to
     be assigned as a second meaning to the same `\( ... \)' construct
     because in practice there is no conflict between the two meanings.
     Here is an explanation:

`\DIGIT'
     after the end of a `\( ... \)' construct, the matcher remembers the
     beginning and end of the text matched by that construct.  Then,
     later on in the regular expression, you can use `\' followed by
     DIGIT to mean "match the same text matched the DIGIT'th time by the
     `\( ... \)' construct."

     The strings matching the first nine `\( ... \)' constructs
     appearing in a regular expression are assigned numbers 1 through 9
     in order that the open-parentheses appear in the regular
     expression.  `\1' through `\9' may be used to refer to the text
     matched by the corresponding `\( ... \)' construct.

     For example, `\(.*\)\1' matches any newline-free string that is
     composed of two identical halves.  The `\(.*\)' matches the first
     half, which may be anything, but the `\1' that follows must match
     the same exact text.



Fri, 15 May 1998 03:00:00 GMT  
 Regular expressions, pattern matching

Quote:

> The Python library documentation says that pattern matching is done like
> Emacs, and as far as I can see, gives no further explanation. I don't have the
> faintest idea how Emacs pattern matching works, or how the various pattern
> matching modules/functions interact, so can someone point me to something like
> the extensive Perl documentation.

This is copied w/o permission from O'Reilly's perl book, p. 25

        .       matches any character except newline
 [a-z0-9]       matches any single character of set
 [^a-z0-9]      matches any single character NOT in set
       \d       matches a digit, same as [0-9]
       \D       matches a non-digit, same as [^0-9]
       \w       matches an alphanumeric (word) character [a-zA-Z0-9_]
       \W       matches a non-word character [^a-zA-Z0-9_]
       \s       matches a whitespace char (space, tab, newline...)
       \S       matches a non-whitespace char
       \n       matches a newline
       \r       matches a return
       \t       matches a tab
       \f       matches a formfeed
       \b       matches a backspace (inside [] only)
       \0       matches a null char
     \000       also matches a null char because...
     \nnn       matches an ASCII char of that octal value
     \xnn       matches an ASCII char of that hex value
     \cX        matches an ASCII control char
 \metachar      matches the char itself (\|,\.,\*...)
    (abc)       remembers the match for later backreferences
      \1        matches the first set of parens matched
      \2        matches whatever the second set of parens matched
      \3        and so on...

      x?        matches 0 or 1 x's, where x is any of the above
      x*        matches 0 or more x's
      x+        matches 1 or more x's
   x{m,n}       matches at least m x's but no more than n

   a,b,c        matches all of a, b, and c in order
fee|fie|foe     matches one of fee, fie, or foe

      \b        matches a word boundary (outside [] only)
      \B        matches a non-word boundary
       ^        anchors match to the beginning of a line or string
       $        anchors match to the end of a line or string

                                                                Dave



Fri, 15 May 1998 03:00:00 GMT  
 Regular expressions, pattern matching

    SM> I'm not sure how standard the syntax selection features are.
    SM> I've personally never used them before, so it wouldn't bother
    SM> me to give them up. I'm sure others would wail at the loss of
    SM> flexibility should they not be present in another regular
    SM> expression package.

I'm pretty sure the GNU regex C library has syntax selection, but
there's the sticky problem of the GPL.  What other regex libraries
might we use?

-Barry



Sat, 16 May 1998 03:00:00 GMT  
 Regular expressions, pattern matching
While we're on the subject, does anybody know where I can
find a formal definition of "regular expressions"? I mean,
all the things I've seen out there that are called regular
expressions have quite a bit in common both in terms of
what they do and what notation they use -- but what makes
one tool "regular expressions" and another just generically
"pattern matching"? I'm not up on the history of this stuff.

------------------------------------------------
David Moles               Phone: +81 3-3720-1640
3-13-7-101 Kita-senzoku, Ota-ku, Tokyo 145 Japan
     http://www.eccosys.com/PEOPLE/deivu/
------------------------------------------------



Sat, 16 May 1998 03:00:00 GMT  
 Regular expressions, pattern matching

Quote:


>>  The Python library documentation says that pattern matching is
>> done like Emacs, and as far as I can see, gives no further
>> explanation. I don't have the faintest idea how Emacs pattern
>> matching works, or how the various pattern matching
>> modules/functions interact, so can someone point me to something
>> like the extensive Perl documentation.

Dave> This is copied w/o permission from O'Reilly's perl book, p. 25

Unfortunately, the \s is not supported in python - I still wonder why,
as this is one of the most valuable tokens.

Karsten
--

Aalborg University, Denmark              I Fredrik Bajers vej 7, A1-207
Institute for Electronic Systems         I DK-9220 Aalborg East
http://www.kom.auc.dk/~karthy            I Phone: (+45) 98158522, ext. 4893



Sun, 17 May 1998 03:00:00 GMT  
 Regular expressions, pattern matching
: While we're on the subject, does anybody know where I can
: find a formal definition of "regular expressions"?

You can find an extensive and quite fulfilling definition of "regular
expression" in "Principles of Compiler Design", by Alfred Aho and Jeffrey
Ullman.  This is popularly known as the "Dragon Book".   My copy is
copyright 1977; I think that it has been updated since then.  [Aho and
Ullman were influential researchers in parsers and compilers (Aho is the A
in awk; Ullman had his hands in the Portable C Compiler).]

In short, the authors state that a regular expression is a language.  The
alphabet of the language consists of those ordinary and matching
characters discussed amply in other postings here.

The interesting and slightly subtle point is that each regular expression
is a complete language.

For example "x" is a regular expression defining a language which allows
only one string: "x".  Likewise, "c.*t" defines a language which allows
all strings starting with "c" and ending with "t"...

--



Mon, 18 May 1998 03:00:00 GMT  
 
 [ 7 post ] 

 Relevant Pages 

1. iss-matching - the free Regular Expression / Pattern Matching cluster

2. Bug in regular expression pattern matching?

3. Regular Expression Pattern Matching "State" Object

4. Regular Expression for Match Pattern (string) Function

5. Regular expression string pattern matching: Embedding pop-11 procedures, and more

6. Pattern-matching regular-expression algorithm?

7. pattern matching using regular expressions

8. HELP !!! I want to DISABLE regular expression pattern matching !!! THANKS

9. HELP !!! I want to DISABLE regular expression pattern matching

10. regular expression matching in J ? (or APL)

11. Regular Expression to match HTML elements

12. regular expression: matching ( )

 

 
Powered by phpBB® Forum Software