:TITLE: String comparisons - compare match first last wordend
;#
;# RCSID: $Header: /cvsroot/tcl/tcltutorial/original/Tcl18.lsn,v 1.1 2004/11/04 16:01:13 davidw Exp $
;# Copyright (c) 1995 Clif Flynt
;# 9300 Fleming Rd.
;# Dexter, MI 48130
;# clif@cflynt.com
;# See file "NOTICE" for licensing terms.
;#
:LESSON_TEXT_START_LEVEL 0:
There are 6 string subcommands that do pattern and string matching.
- string compare
string1 string2
- Compares string1 to string2 and returns
- -1 ..... If string1 is less than string2
- 0 ........ If string1 is equal to string2
- 1 ........ If string1 is greater than string2
These comparisons are done lexicographically, not numerically.
- string first
string1 string2
- Returns the index of the character in string1
that starts the first match to string2, or -1 if there is no
match to string2 in string1
- string last
string1 string2
- Returns the index of the character in string1
that starts the last match to string2, or -1 if there is no
match to string2 in string1
- string wordend
string index
- Returns the index of the character just after the last one in the
word which contains the index'th character of string. A word is
any contiguous set of letters, numbers or underscore characters, or a
single other character.
- string wordstart
string index
- Returns the index of the character just before the first one in the
word which contains the index'th character of string. A word is
any contiguous set of letters, numbers or underscore characters, or a
single other character.
- string match
pattern string
- Returns 1 if the pattern matches string. Pattern
is a glob style pattern.
Globbing is the wildcarding technique that the shell uses.
globbing wildcards are:
- *
- Matches any quantity of any character
- ?
- Matches one occurrence of any character
- \X
- The backslash escapes a special character in globbing
just the way it does in Tcl substitutions. Using the backslash lets you use
glob to match a * or ?.
- [...]
- Matches one occurrence of any character within the brackets.
A range of characters can be matched by using a range between the brackets.
For example, [a-z] will match any lower case letter.
:TEXT_END:
:LESSON_TEXT_START_LEVEL 1:
There are 6 string subcommands that do pattern and string matching.
- string compare
string1 string2
- Compares string1 to string2 and returns
- -1 ..... If string1 is less than string2
- 0 ........ If string1 is equal to string2
- 1 ........ If string1 is greater than string2
These comparisons are done lexicographically, not numerically.
- string first
string1 string2
- Returns the index of the character in string1
that starts the first match to string2, or -1 if there is no
match to string2 in string1
- string last
string1 string2
- Returns the index of the character in string1
that starts the last match to string2, or -1 if there is no
match to string2 in string1
- string wordend
string index
- Returns the index of the character just after the last one in the
word which contains the index'th character of string. A word is
any contiguous set of letters, numbers or underscore characters, or a
single other character.
- string wordstart
string index
- Returns the index of the character just before the first one in the
word which contains the index'th character of string. A word is
any contiguous set of letters, numbers or underscore characters, or a
single other character.
- string match
pattern string
- Returns 1 if the pattern matches string. Pattern
is a glob style pattern.
Globbing is the wildcarding technique that the shell uses.
globbing wildcards are:
- *
- Matches any quantity of any character
- ?
- Matches one occurrence of any character
- \X
- The backslash escapes a special character in globbing
just the way it does in Tcl substitutions. Using the backslash lets you use
glob to match a * or ?.
- [...]
- Matches one occurrence of any character within the brackets.
A range of characters can be matched by using a range between the brackets.
For example, [a-z] will match any lower case letter.
:TEXT_END:
:LESSON_TEXT_START_LEVEL 2:
Operations that are often performed on strings include searching for a
pattern, or comparing one string to another. In this lesson the
string commands for searching and comparing are discussed.
Tcl has several methods of searching for one string within another
string. First and last are the string subcommands that will return
the first and last occurrence of one string within another. You
might use these commands to search a text file for a particular word.
The match command will tell you if a pattern occurs within a string.
It uses the same globbing rules as the lsearch command.
The string compare will tell you if one string comes before the other
alphabetically. This is useful to sort a set of strings into alphabetic
order.
The descriptions of these commands is:
- string compare
string1 string2
- Compares string1 to string2 and returns
- -1 ..... If string1 is less than string2
- 0 ........ If string1 is equal to string2
- 1 ........ If string1 is greater than string2
These comparisons are done lexicographically, not numerically.
- string first
string1 string2
- Returns the index of the character in string1
that starts the first match to string2, or -1 if there is no
match to string2 in string1
- string last
string1 string2
- Returns the index of the character in string1
that starts the last match to string2, or -1 if there is no
match to string2 in string1
- string wordend
string index
- Returns the index of the character just after the last one in the
word which contains the index'th character of string. A word is
any contiguous set of letters, numbers or underscore characters, or a
single other character.
- string wordstart
string index
- Returns the index of the character just before the first one in the
word which contains the index'th character of string. A word is
any contiguous set of letters, numbers or underscore characters, or a
single other character.
- string match
pattern string
- Returns 1 if the pattern matches string. Pattern
is a glob style pattern.
Globbing is the wildcarding technique that the shell uses.
globbing wildcards are:
- *
- Matches any quantity of any character
- ?
- Matches one occurrence of any character
- \X
- The backslash escapes a special character in globbing
just the way it does in Tcl substitutions. Using the backslash lets you use
glob to match a * or ?.
- [...]
- Matches one occurrence of any character within the brackets.
A range of characters can be matched by using a range between the brackets.
For example, [a-z] will match any lower case letter.
In the example, first and last are used to
find the first and last slash in a path. By knowing if the first character
is a "/", we can tell if a path is absolute or relative.
The value of
last is incremented before we test to see if the "/" is
the last character in the string because Tcl numbers characters in a
string from 0. If there were two characters in a string, string length
would return a 2, obviously, and string last would report the
last character at position 1. If the "/" is the last character, then
last is moved back 2 spaces, since string range
returns the all the characters from start to end, including the start and
end positions.
In the final section of code, the comparison to "a", we determine
whether the first character of the string is upper or lower case. The
order of the characters that a computer uses go from 0 to 9, then A
Z, and finally a to z. Any letter less than 'a' is an upper case
letter. This is a very simple minded test that will report numbers as
being uppercase.
Try rewriting this example using the split command to separate
the parts of the path into separate list elements.
:TEXT_END:
:CODE_START:
set fullpath "/usr/home/clif/TCL_STUFF/TclTutor/Lsn.17"
set relativepath "CVS/Entries"
set directorypath "/usr/bin/"
set paths [list $fullpath $relativepath $directorypath]
foreach path $paths {
set first [string first "/" $path];
set last [string last "/" $path];
;# Report whether path is absolute or relative
if {$first != 0} {
puts "$path is a relative path"
} else {
puts "$path is an absolute path"
}
;# If "/" is not the last character in $path, report the last word.
;# else, remove the last "/", and find the next to last "/", and
;# report the last word.
incr last;
if {$last != [string length $path]} {
set name [string range $path $last end];
puts "The file referenced in $path is $name"
} else {
incr last -2;
set tmp [string range $path 0 $last];
set last [string last "/" $tmp];
incr last;
set name [string range $tmp $last end]
puts "The final directory in $path is $name"
}
;# CVS is a directory created by the CVS source code control system.
;#
if {[string match "*CVS*" $path]} {
puts "$path is part of the source code control tree"
}
;# Compare to "a" to determine whether the first char is upper or lower case
set comparison [string compare $name "a"]
if {$comparison >= 0} {
puts "$name starts with a lowercase letter\n"
} else {
puts "$name starts with an uppercase letter\n"
}
}
:TEXT_END: