The REGEXP_SUBSTR scalar function returns one occurrence of a substring of a string that matches the regular expression pattern.

The schema is SYSIBM.

source-string

An expression that specifies the string in which the search is to take place.
This expression must return a built-in character string, graphic string, numeric value, Boolean value, or datetime value.
A numeric, Boolean, or datetime value is implicitly cast to VARCHAR before the function is evaluated.
A character string cannot specify the FOR BIT DATA attribute (SQLSTATE 42815).

pattern-expression

An expression that specifies the regular expression string that is the pattern for the search.
This expression must return a built-in character string, graphic string, numeric value, Boolean value, or datetime value.
A numeric, Boolean, or datetime value is implicitly cast to VARCHAR before the function is evaluated.
The length of a CLOB or DBCLOB expression must not be greater than the maximum length of a VARCHAR or VARGRAPHIC data type.
A character string cannot specify the FOR BIT DATA attribute (SQLSTATE 42815).

start

An expression that specifies the position within source-string at which the search is to start.
The expression must return a built-in character string, graphic string, Boolean, or numeric value.
If the value is not of type INTEGER, it is implicitly cast to INTEGER before the function is evaluated.
The value of the integer must be greater than or equal to 1.
If OCTETS is specified and the source string is graphic data, the value of the integer must be odd (SQLSTATE 428GC).
The default start value is 1.
See parameter description for CODEUNITS16, CODEUNITS32, or OCTETS for the string unit that applies to the start position.

occurrence

An expression that specifies which occurrence of the pattern expression within source-string to search for.
The expression must return a built-in character string, graphic string, Boolean, or numeric value.
If the value is not of type INTEGER, it is implicitly cast to INTEGER before the function is evaluated.
The occurrence value must be greater than or equal 1.
The default occurrence value is 1, which indicates that only the first occurrence of the pattern expression is considered.

flags

An expression that specifies flags that controls aspects of the pattern matching.
The expression must return a built-in character string that does not specify the FOR BIT DATA attribute (SQLSTATE 42815).
The string can include one or more valid flag values and the combination of flag values must be valid (SQLSTATE 2201T).
An empty string is the same as the value 'c'. The default flag value is 'c'.

Table 1. Supported flag values

Flag value	Description
c	Specifies that matching is case-sensitive. This flag is the default value if 'c' or 'i' is not specified. This value must not be specified with a value of 'i'.
i	Specifies that matching is case insensitive. This value must not be specified with a value of 'c'.
m	Specifies that the input data can contain more than one line. By default, the '^' in a pattern matches only the start of the input string; the '$' in a pattern matches only the end of the input string. If this flag is set, "^" and "$" also matches at the start and end of each line within the input string.
n	Specifies that the '.' character in a pattern matches a line terminator in the input string. By default, the '.' character in a pattern does not match a line terminator. A carriage-return and line-feed pair in the input string behaves as a single-line terminator, and matches a single "." in a pattern.
s	Specifies that the '.' character in a pattern matches a line terminator in the input string. This value is a synonym for the 'n' value.
x	Specifies that white space characters in a pattern are ignored, unless escaped.

group

An expression that specifies which capture group of the pattern expression within source string to return.
The expression must return a built-in character, binary, or graphic string, or a Boolean value.
If the value is not of type INTEGER, it is implicitly cast to INTEGER before the function is evaluated.
The group value must be greater than or equal to 0 and must not be greater than the number of capture groups in the pattern expression (SQLSTATE 22546).
The default group value is 0, which indicates that the string that matches the entire pattern is to be returned.

CODEUNITS16, CODEUNITS32, or OCTETS

Specifies the string unit of the start value:

• CODEUNITS16 specifies that the start value is expressed in 16-bit UTF-16 code units.
• CODEUNITS32 specifies that the start value is expressed in 32-bit UTF-32 code units.
  This is the default.
• OCTETS specifies that the start value is expressed in bytes.

If the string unit is specified as CODEUNITS16 or OCTETS, and if the string unit of the source string is CODEUNITS32, an error is returned (SQLSTATE 428GC).

For more information, see "String units in built-in functions" in Character strings.

Result

The result of the function is a string.
The data type of the string is the same data type as the source string, except for CHAR, which becomes VARCHAR; and GRAPHIC, which becomes and VARGRAPHIC.
The length attribute of the result data type is same as the length attribute of the source string.
The actual length of the result is the length of the occurrence in the string that matches the pattern expression.
If the pattern expression is not found, the result is the null value.

The result of the REGEXP_SUBSTR function can be null. If any argument is null, the result is the null value.

Notes

• The regular expression processing is done by using the International Components for Unicode (ICU) regular expression interface.
• Considerations for non-Unicode databases:
  • A regular expression pattern supports only halfwidth control characters; use a character string data type for the pattern expression argument.
    A character string data type can be used for the pattern expression argument even when a graphic string data type is used for the source string argument.
  • The source string argument must be a graphic string data type if the pattern expression argument is a graphic string data type.

Examples

1. Return the string which matches any character preceding a 'o'.

      SELECT REGEXP_SUBSTR('hello to you', '.o',1,1) 
      FROM   SYSIBM.SYSDUMMY1

   The result is 'lo'.
2. Return the second string occurrence which matches any character preceding a 'o'.

      SELECT REGEXP_SUBSTR('hello to you', '.o',1,2) 
      FROM   SYSIBM.SYSDUMMY1

   The result is 'to'.
3. Return the third string occurrence which matches any character preceding a 'o'.

      SELECT REGEXP_SUBSTR('hello to you', '.o',1,3) 
      FROM   SYSIBM.SYSDUMMY1

   The result is 'yo'.