Just FYI, Twitter screen names are not (or, apparently, didn't use to be) restricted to 0-9A-Z-_
<id>6295462</id> <screen_name>Magic carpet1</screen_name> <id>3939231</id> <screen_name>Walking the dog2</screen_name> <id>92595586</id> <screen_name>Its mine\nM2</screen_name> We're also seeing non-ASCII in some other screen names. (Though we don't currently know what they are... we just know they exist! ;) We're assuming this is a vestige of long ago coding? It does screw up our desire to use ascii char(15) in db... Maybe Twitter can fix?