Best way to replace characters

Question

Hi
Im having some data corruption on the database where the german letters have been replaced by certain symbols.
These letters have been replaced by:
ü = }
ä = {
ö = |
So instead of Grübbestrasse, its Gr}bbestrasse.
I now want update alot of this data but Im considering which way is the best to go.
One way which Iv considered (but I dont think is the best) is this:
I substring the address from start pos 0 to pos where the "}" ends (then minus 1) - concat with the real letter - subString where the "}" starts and where the column ends.
update address_table
set address_1 = subStr(address1,0,(instr(address1,'}')-1)) ||'ü'|| subStr(address1,(inStr(address11,'}')+1),20)
Do you have any better ideas ?
Because I know this will not work if I have more then 1 wrong character in the same address.
Cheers

Answer 1

Hi,
You can use the replace function to perform that:
UPDATE address_table
   SET address_1 = replace(address1, '}', 'ü');or the translate to run all those in a single pass:
with address_table as (
select 'Gr}bbestrasse {|' address_1 from dual)
select replace(address_1, '}', 'ü'),
       translate(address_1, '}{|', 'üäö')
  from address_tableYou just need to be sure there won't be false-positives where the '}' character was correctly there.
[ Docs on the Replace function|http://download.oracle.com/docs/cd/E11882_01/server.112/e10592/functions151.htm#SQLRF00697]
Regards
Edited by: fsitja on Jan 27, 2010 3:45 PM edited as Michaels did it too. Too bad I have no idea what those mean... had to Google it to try and figure out. :)