<div class="moz-cite-prefix">On 05/19/2014 01:54 AM, Andreas wrote:<br /></div><blockquote
cite="mid:5379C6C1.5020904@gmx.net"type="cite"><font face="Arial">Hi<br /><br /> I need to clean up phone-numbers.
SomehowI got a Excel list that has weird graphical characters trailing some of the entries.</font><br /><font
face="Arial"><fontface="Arial">My DB is UTF8 so it would store this mess but I don<small>'t like to import it in the
firstplace.<br /><br /> OK, I know how to read the stuff into a temporary table to clean it up before the actual
import.<br/></small></font>How can I do an update on the column that deletes every char that is not in a given set of
charslike '+- 0123456/()'?<br /><br /></font></blockquote><br /> See: <a class="moz-txt-link-freetext"
href="http://www.postgresql.org/docs/current/static/functions-matching.html">http://www.postgresql.org/docs/current/static/functions-matching.html</a><br
/><br/> For the first case, the regexp_replace function is probably your best bet. But note that, depending on the
qualityof your input, just removing characters outside that range may still not yield the desired result.<br /><br />
selectregexp_replace('(12s3)-456-635/6(a+sdk', '[^0-9()+-/]', '', 'g');<br /> regexp_replace <br />
-------------------<br/> (123)-456-635/6(+<br /><br /> You can remove all formatting by requiring only digits then
checkand/or reformat later as desired.<br /> steve=> select regexp_replace('(12s3)-456-6356(a+sdk', '[^0-9]', '',
'g');<br/> regexp_replace <br /> ----------------<br /> 1234566356<br /><br /><blockquote
cite="mid:5379C6C1.5020904@gmx.net"type="cite"><font face="Arial"><br /> Second but similar question:<br /> How can I
selectrecords that have fields that contain characters not included in a given alphabet?<br /> E.G. find fields that
containsome char not in 0-9,a-z,A-Z, +-()/?<br /></font><br /></blockquote> See regexp_match on the above-referenced
page.<br/><br /> Cheers,<br /> Steve<br /><br />