CSVs: Watch your language!

timallen

timallen

Well Known Member
We recently had a strange problem with running Universal Batch Engine (UBE) reports on the Enterprise Server. When run with the Comma Separated Values (CSV) option, they came out differently on the Enterprise Server than when they were run locally.

We are in Spain. This will become important shortly.

When the client told me about the problem, I thought what I always do: PEBCAK. But as it turns out, the output on the server *was* different from the local output. I tried it myself, and sure enough, the client was right! I was stumped.

Then I looked at some of the numbers being produced: LOCAL: 123.121,34. 65.212,21. SERVER: 123123.34. 65212.21. Aha.

I went to the regional settings on the Enterprise Server: American. The regional settings on the client: Spanish (remember what I was saying earlier?). So I had a choice: change regional settings on 120 client machines, or change regional settings on one Enterprise Server. The Enterprise Server it was. Once I changed regional settings on the Enterprise Server to Spanish, everything started working.

One strange thing. With the regional setting on "Spanish" on the Enterprise Server, the output really isn't CSV. It's Tab-Separated Values. And the file comes out named "whatever.TXT". With regional settings set to "American", the output really is CSV, and it's named "whatever.CSV". Looks like they didn't quite finish this part. JDE-FU.
 
In the server's [UBE] section, try this:

prtCSVExtension=.csv
prtCSVSeparator=,

Worked for me. I had to get this info about a year ago. If you aren't on the right sevice pack, it may cause problems for you, but I'm pretty sure that from about sp19 on it works great.
 
Excellent! I did not know that it had those settings-- since everyone here has this problem, I guess I'll need to set this in all clients. Thanks.
 
Back
Top