Re: ANSI-characters with Informix ?

> Martin.Berns@Materna.DE (Martin Berns) writes: > In article <44e3fg$40p@siemens.co.at>, > helmut.zoerrer@banyan.siemens.co.at says... > > > >We've upgraded from Informix 6.0 to Informix 7.1. > >Now if i'm tryig to insert special Windows-Characters like <ae> (email > >is also 7 Bit) using dbaccess > >INSERT int tab (attr1, attr2) values ("Hallo","H<a>llo") > >where <a> stands for the special character ae > > > >I get the error ''202 an illegal character has been found in the > >statement" > [lines deleted] > hallo Helmut, > the problem you describe looks like you did not activate the NLS-function > before creating your database > [lines deleted] This looked frightening so I tested it, and got the same problem. You can insert these characters via a standard 4GL program without setting any NLS-functions but dbaccess complains. Even without NLS used in any way the database should support 8 bit characters and I see this behaviour as a clear defect that Informix must fix. This is an 8 bit character within a quoted string, *not* as part of a name or anything else. I have read all about NLS in the Informix manuals, but not realy studied it yet. I never found any explanation on what is actually going on inside the database when it comes to indexing, sorting and so on, but there where warnings about possible performance problems. As I understand it you are also supposed to start using nchar in stead of simply char for all your db-fields and variable definitions in 4GL (and NewEra). That is an impossible task for us, so we desided to skip it, and simply run standard US setup assuming 8 bit characters where supported fully. We realize the problems we get with sorting, but they are not that significant for us yet. We also have to maintain our programs for users of SE ver. 5.x with no NLS support. And the extra problems of all sorts with setting up NLS in Unix and finding out how it works, if at all, is not something we would love to do. We have other things to spend time on. The whole enchelada is so incredibly complex we flinch at even starting when all we realy would need is a different sort order for indexed reads and upper/lowe case functions that new about the whole ISO8859 (same as MS Windows) character set. Comments from anyone who have used NLS would be appreciated. Nils.Myklebust@ccmail.telemax.no NM-data, Dalsbergstien 7, N-0170 Oslo, Norway My opinions are those of my company