ABAP Keyword Documentation → ABAP − Reference → Processing Internal Data → Character String and Byte String Processing → Expressions and Functions for String Processing → Regular Expressions → Syntax of Regular Expressions
Single Character Patterns
Single characters are represented by literal characters or operators. If preceded by a backslash \, a special character of an operator is interpreted as a literal character. This applies in particular for the backslash \ itself, so that the regular expression \ is the same as the single character \. If the backslash is followed by a literal character, the backslash is ignored as if it does not exist.
Other versions: 7.31 | 7.40 | 7.54
Literal Characters
A literal character is a character that is not a special character, or a special character preceded by a backslash \, or enclosed between \Q ... \E. As a search string, a literal character matches the same single character exactly.
Note
Case-sensitivity can be controlled using the appropriate commands or methods.
Examples
The following table shows some results of a test that is case-sensitive.
| Pattern | text | match |
|---|---|---|
| A | A | X |
| A | a | - |
| . | . | X |
| A | AB | - |
| AB | AB | X |
The regular expression AB is a concatenation of two expressions for single characters.
Operators for Single Characters
These operators are made up of the special characters ., [, ], ^, and -. The last two are only effective as special characters in specific positions within [ ]. The special characters can be made into literal characters using the prefix \.
Placeholders for Single Characters
The special character . is a placeholder for any single character. The operator \C has the same effect as the special character .. A regular expression . or \C matches exactly one single character.
Examples
The following table shows some results of a test that is not case-sensitive.
| Pattern | Text | match |
|---|---|---|
| . | A | X |
| \C | a | X |
| . | AB | - |
| .. | AB | X |
The regular expression .. is a concatenation of two expressions for single characters.
Self-Defined Sets for Single Characters
The special characters [ ] can be placed around any number of literal characters or names for character classes (see below), and thus define a set of literal characters. A regular expression [...] matches exactly one single character specified as a literal character within the brackets, or which is contained in a specified character class. At least one literal character or one name for a character class (see below) must be contained within the brackets. A character [ or ], which is positioned directly after the opening bracket, is interpreted as a literal character. Some of the special characters that start with a backslash, such as \A or \Q, lose their special function within sets, and are interpreted as the simple literal character A or Q.
Examples
The following table shows some results from a test.
| Pattern | Text | match |
|---|---|---|
| [ABC] | B | X |
| [ABC] | ABC | - |
| [AB][CD] | AD | X |
| [\d] | 9 | X |
The regular expression [AB][CD] is a concatenation of two expressions for single characters.
Negation of a Self-Defined Set for Single Characters
If the character ^ is the first character in a self-defined set for single characters and is specified directly after [, it acts as a special character and negates the rest of the set of literal characters or character classes. A regular expression [^...] matches precisely one single character that is not specified as a literal character within the square brackets or that is not in a specified character class. A character ^ that is not specified directly after [ acts as a literal character.
Examples
The following table shows some results from a test.
| Pattern | Text | match |
|---|---|---|
| [^ABC] | B | - |
| [^ABC] | Y | X |
| [^A][^B] | BA | X |
| [A^B] | ^ | X |
The regular expression [^A][^B] is a concatenation of two expressions for single characters.
Ranges in a Self-Defined Set for Single Characters
If the character - is between two literal characters, it acts as a special character and defines a range between the literal characters. The range is the set of characters that is enclosed by literal characters in the code page of the current operating system. A regular expression [...-...] matches exactly one single character that is within the defined range. A character -, which is not between two literal characters, acts as a literal character. A literal character cannot be part of two ranges, for example, 'a-z-Z' is not a regular expression.
Examples
The following table shows some results from a test.
| Pattern | Text | match |
|---|---|---|
| [A-Za-z0-9] | B | X |
| [A-Za-z0-9] | 5 | X |
| [A-Za-z0-9] | # | - |
| [A-Za-z0-9] | - | - |
| [A-Za-z0-9-] | - | X |
In the last expression, the closing - does not act as a special character.
Character Classes
Within sets for single characters defined using [ ], predefined character classes can be specified for certain sets for single characters whose behavior can, however, depend on the language and platform.
- [:alnum:]
Set of all alphanumeric characters (union of [:alpha:] and [:digit:])
- [:alpha:]
Set of all uppercase and lowercase letters including language-specific special characters (umlauts, accents, diphthongs) without digits.
- [:blank:]
Blank characters and horizontal tabs
- [:cntrl:]
Set of all control characters
- [:digit:]
Set of all digits 0 to 9
- [:graph:]
Set of all displayable characters except for blanks and horizontal tabs
- [:lower:]
Set of all lowercase letters including language-dependent special characters (umlauts, accents, diphthongs)
- [:print:]
Set of all displayable characters (union of [:graph:] and [:blank:])
- [:punct:]
Set of all punctuation characters
- [:space:]
Set of all blank characters, tabs, and carriage feeds
- [:unicode:]
Set of all characters whose character representation is greater than 255
- [:upper:]
Set of all uppercase letters including language-dependent special characters (umlauts, accents, diphthongs)
- [:word:]
Set of all alphanumeric characters including underscore _
- [:xdigit:]
Set of all hexadecimal digits ("0"-"9", "A"-"F", and "a"-"f")
Notes
- Character classes only act within [ ] as specified. A regular expression [:digit:] does not define the set of all digits, but instead defines a character set consisting of ":", "d", "g", "i", and "t". To specify the set of all digits, use the regular expression [[:digit:]].
- Due to their dependencies on language and platform, these character classes must be used with care.
Examples
The following table shows some results of a test that is case-sensitive.
| Pattern | Text | match |
|---|---|---|
| [[:alnum:]] | a | X |
| [[:alnum:]] | ; | - |
| [[:alpha:]] | 1 | - |
| [[:punct:]] | . | X |
| [[:lower:]] | â | X |
| [[:upper:]] | Ä | X |
| [[:digit:][:alpha:]]+ | a1b2c3 | X |
Abbreviations for Character Sets
For frequently used character sets, specific operators are available as abbreviations:
| Character Set | Abbr. | Meaning |
|---|---|---|
| [[:digit:]] | \d | Placeholder for a digit |
| [^[:digit:]] | \D | Placeholder for a non-digit |
| [[:lower:]] | \l | Placeholder for a lowercase letter |
| [^[:lower:]] | \L | Placeholder for a character that is not a lowercase letter |
| [[:space:]] | \s | Placeholder for a blank character |
| [^[:space:]] | \S | Placeholder for a non-blank character |
| [[:upper:]] | \u | Placeholder for an uppercase letter |
| [^[:upper:]] | \U | Placeholder for a character that is not an uppercase letter |
| [[:word:]] | \w | Placeholder for an alphanumeric character plus underscore _ |
| [^[:word:]] | \W | Placeholder for an non-alphanumeric character without underscore _ |
Note
If case is ignored by the ABAP statements FIND and REPLACE
and when creating an object of the class CL_ABAP_REGEX, \l and \u
are equivalent to [[:alpha:]] or \L and
\U is equivalent to [^[:alpha:]]. The special
characters \w, \u, \l,
\d, and \s can also be specified within
sets [...]. Use of the special characters
\W, \U, \L,
\D, and \S within sets is not permitted and raises the exception CX_SY_INVALID_REGEX.
Examples
The following table shows some results of a test that is case-sensitive.
| Pattern | Text | match |
|---|---|---|
| \d | 4 | X |
| \D | ; | X |
| \l | u | X |
| \l | U | - |
| \L | S | X |
| \s | X | |
| \S | # | X |
| \u | U | X |
| \U | . | X |
| \w | A | X |
| \w | 8 | X |
| \W | : | X |
| \W | _ | - |
Equivalence Classes
The operators [..] and [==] are reserved for future language enhancements and currently raise the exception CX_SY_INVALID_REGEX if used in sets.