Dealing With Special Characters ...

bmouton · Post by **bmouton** » Sat Mar 28, 2009 7:14 am

I'm running into a strange issue ... I'm running Datastage Server v8, DB2 9.5 on SUSE Linux.

Here's the deal:
I have a person's name that contains the O with two dots on top (sorry don't know the name of that character). I can insert the name with no issue. I can substring out the first character into a hash file (the O with the two dots on top). However, when I try to load the single character into the database, it rejects the reject.

We are running DB2 9.5 UTF-8. The field was character, but we changed it ot varchar(8) thinking that is was trying to take up more space in the field.

I have tried None, UTF-8 for the NLS and nothing works ....

Please help!!!

chulett · Post by **chulett** » Sat Mar 28, 2009 8:42 am

Sometimes known as an "O-with-dots", the dots are a diacritic mark known as an umlaut or diaeresis or trema as well, it seems. So "O-umlaut".

Is your install "NLS enabled"? Assuming yes, but need to double-check. Also what stage are you using with DB2?

chulett · Post by **chulett** » Sat Mar 28, 2009 8:43 am

[nothing to see here, move along]

bmouton · Post by **bmouton** » Sat Mar 28, 2009 9:05 am

[quote="chulett"]Sometimes known as an "O-with-dots", the dots are a diacritic mark known as an [url=http://en.wikipedia.org/wiki/%C3%96]umlaut[/url] or diaeresis or trema as well, it seems. So "O-umlaut". :wink:

Is your install "NLS enabled"? Assuming yes, but need to double-check. Also what stage are you using with DB2?[/quote]

Thanks for the definition of that character ...

NLS is enabled ... We are using the DB2 Stage (but it's supposed to only be compatible with DB2 8.x ... We're running 9.5 ... I'm wondering if making that field larger will solve the problem ... I'm not sure how many characters those type of fields take up ... We have no problem taking a name with special characters (fields are varchar(128) ...

We are opening a ticket with IBM on Monday.

chulett · Post by **chulett** » Sat Mar 28, 2009 9:19 am

bmouton wrote:However, when I try to load the single character into the database, it rejects the reject.

Rejects how, exactly? Can you post the entire message?

Yes, UTF-8 is a "multi-byte encoding" character set. Certain things like your umlaut can cause it to need two (and other things even three) bytes to store a single character. I don't know DB2, but in Oracle I can define a CHAR(1) field as being "one" byte or character in length and it does make a difference for stuff like this. And sometimes you can get away with leaving the db target alone and simply increasing the size of the metadata for that field in the job itself, you might want to try that first.

chulett · Post by **chulett** » Sat Mar 28, 2009 9:21 am

If you're going to quote people or otherwise make use of the [tags] that are available here, make sure you do not check the "Disable BBCode in this post" option or they won't work... as you can see in your reply.

bmouton · Post by **bmouton** » Sat Mar 28, 2009 9:24 am

I really think it's a DB2 thing ... My experience with Teradata, Oracle, Ingres, and SQL Server using other ETL tools (e.g. Ab Initio, Informatica) have never had this issue ....

DSXchange

Dealing With Special Characters ...

Dealing With Special Characters ...

Re: Dealing With Special Characters ...