Equivalent to regexp_replace((^[A-Z])|([IJ])|(?<= ).) in regexp

In DB2, the next query works fine:

select  REGEXP_REPLACE( 'René de la Ömer-Arie IJsbrand IJsse'
                      ,'(^[A-Z])|([IJ])|(?<= ).','') FROM "SYSIBM".SYSDUMMY1

Result: ené e a mer-Arie sbrand sse

However, I would like to achieve the opposite result:

RdlÖIJIJ

If I try the regexp_substr equivalent, it doesn't give me the desired outcome:

select  REGEXP_SUBSTR( 'René de la Ömer-Arie IJsbrand IJsse'
                  ,'(^[A-Z])|([IJ])|(?<= ).') FROM "SYSIBM".SYSDUMMY1

Result: R

Is there a way to achieve the result "RdlÖIJIJ" with (or without) regexp_substr?

CodePudding user response：

You could match a char other than I or J while asserting a non whitespace char to the left, and replace with an empty string.

(?<=\S)[^IJ]