RE: Splitting field

Hello Rob. Thanks for your reply. > >Wild idea... Do you have any kind of machine readable street listing >from your postal service? You could cross reference by postal code (ZIP >in the US) and narrow down lines of data to be reviewed manually if the >ST_NAME doesn't match the street listing. That's an good idea to a listing. We don't have an (actual) street listing of germany. > >Just an idea... The pattern matching algorithm would probably be >difficult to write in 4GL unless you used exact match or %STRING% type >matching. That could be work. But I don't use 4GL. My database uses only isql. Dirk >Rob > >-----Original Message----- >From: Dirk Emmermacher [mailto:emmermacher@hotmail.com] >Sent: Monday, April 24, 2006 5:38 AM >To: Konikoff, Rob (Contractor); obnoxio@serendipita.com; >keith.simmons@office2office.biz >Cc: informix-list@iiug.org >Subject: RE: Splitting field > >Hello Rob. > >Thanks for your answer (and all others too). > >I will test it. I hope the result will not be to cruel. >I'm really, really happy about this change of structure... >... I have to modify a really large number of reports and forms too ;-). > >Best regards. > >Dirk > > > > > > -> How can I do this with my old IDS 7.30? > > > -> A streetname can contains numbers and some house numbers have > > > -> characters... > > > >1. Unload your KEY and your ADDRESS column > >2. Split the column based on white space, probably using AWK outside >the > >database, or an offline standalone package that supports space > >delimiters, depending on the size/amount of records > >3. ALTER table to include your new columns, ST_NUM and ST_NAME > >4. Reload your new columns based on the KEY and plug in the new data > >5. ALTER table to drop your old column of combined ADDRESS data > >6. Pray your data integrity guys don't kill you > > > >Rob >