Say you have a phone number in various formats but you just want the numeric characters and all the non-numeric characters removed. Well use the SQL function REGEXP_REPLACE and look for characters with the regular expression [^\d] and replace it with empty character ”. [^\d] is a negated set which means match any characters not in between [^ and ]. \d is a short hand way of saying all digit characters 0-9.

SELECT PHONE_NUMBER, REGEXP_REPLACE(PHONE_NUMBER,'[^\d]’,”)
FROM MYTABLE;

123-123-1234,1231231234

References:
IBM documentation on REGEXP_REPLACE
http://www.ibm.com/support/knowledgecenter/ssw_ibm_i_71/db2/rbafzscaregexp_replace.htm

Online regex tester tool
http://regexr.com/