Regular Expressions 4

Specifying WHAT to Match - Part 2

In this tutorial we continue examining ways to specify WHAT text to match by exploring POSIX character classes and equivalence classes.
Character Classes
We saw earlier how a range expression like '[a-z]' works well with the English alphabet but it does not yield accurate results with other alphabets. POSIX character classes, on the other hand, are portable between languages.
With character classes you can match any one of a group of characters, like a lowercase character or a numeric digit, inside a character list pattern. Character class names take the format '[:class:]'. A character list containing a character class looks like '[[:class:]]'. "class" can be any one of the following terms.

Class	Description
lower	lowercase letters
upper	uppercase letters
alpha	all characters in the classes 'lower' and 'upper'
digit	numeric digits
xdigit	hexadecimal digits
alnum	all characters in the classes 'alpha' and 'digit'
blank	and
space	white-space characters
cntrl	control characters
punct	punctuation characters
graph	printable characters, not including
print	printable characters, including

This example shows how the class named 'lower' works, an alternative to using the range '[a-z]'.

execute set_pattern( '[[:lower:]]' )

execute set_target( '+24'   )

execute add_target( '-abc'  )

execute add_target( '.1415' )

execute add_target( '@@@'   )

execute add_target( 'ABC'   )

execute add_target( 'àéîõü' )

select * from test_results ;

PATTERN	TARGET	MATCH	MATCHED_VALUE	POSITION
[[:lower:]]	24	N	(null)	0
[[:lower:]]	-abc	Y	a	2
[[:lower:]]	0.1415	N	(null)	0
[[:lower:]]	@@@	N	(null)	0
[[:lower:]]	ABC	N	(null)	0
[[:lower:]]	…‚Œo	Y	o	4

The 'digit' class includes all numeric digits.

execute set_pattern( '[[:digit:]]' )

select * from test_results ;

PATTERN	TARGET	MATCH	MATCHED_VALUE	POSITION
[[:digit:]]	24	Y	2	2
[[:digit:]]	-abc	N	(null)	0
[[:digit:]]	0.1415	Y	1	2
[[:digit:]]	@@@	N	(null)	0
[[:digit:]]	ABC	N	(null)	0
[[:digit:]]	…‚Œo	N	(null)	0

To get all characters in both the 'lower' and the 'digit' classes, combine them like this.

execute set_pattern( '[[:lower:][:digit:]]' )

column pattern format a25

select * from test_results ;

PATTERN	TARGET	MATCH	MATCHED_VALUE	POSITION
[[:lower:][:digit:]]	24	Y	2	2
[[:lower:][:digit:]]	-abc	Y	a	2
[[:lower:][:digit:]]	0.1415	Y	1	2
[[:lower:][:digit:]]	@@@	N	(null)	0
[[:lower:][:digit:]]	ABC	N	(null)	0
[[:lower:][:digit:]]	…‚Œo	Y	o	4

column pattern format a15

This example shows how to exclude all characters in both classes.

execute set_pattern( '[^[:lower:][:digit:]]' )

column pattern format a25

select * from test_results ;

PATTERN	TARGET	MATCH	MATCHED_VALUE	POSITION
[^[:lower:][:digit:]]	24	Y	+	1
[^[:lower:][:digit:]]	-abc	Y	-	1
[^[:lower:][:digit:]]	0.1415	Y	.	1
[^[:lower:][:digit:]]	@@@	Y	@	1
[^[:lower:][:digit:]]	ABC	Y	A	1
[^[:lower:][:digit:]]	…‚Œo	Y	…	1