DS 12.0 and above - DSCONFIG and UTF8

[system]

I’m confused about whether or not DSCONFIG should be forcibly set to UTF8, like:

AL_Engine=_.utf-8

Manoj recently seemed to say it should be set that way:

120×120 10.4 KB

Hi, We’re consuming a webservice that returns a XML with Spanish characters, such as “Ñ” and “á”. The engine codepage is UTF-16 and we’ve tried running the webservice with the datastore’s encoding UTF-8 and CP1252. We call the webservice in a batch job and it works fine until it processes the very first character in Spanish encoding, at this moment it stops running whatever codepage we use in the webservice datastore. Which codepage should we use? What can we do to process the entire XML re…

But Werner seems to indicate that it should be set to Default only:

120×120 10.4 KB

Hi, We are trying to write data into a SAP BW target. When writing to a demo Master Transfer Structure … We get the error: Load to BW failed: RequestID = <>, Packet Number = <-1> We have a second DF that writes to a Transaction Transfer Structure … This has no errors, but no rows are written to the SAP Target. Can anyone enlighten us? Where are the log files? We are AIX unix DS version 12.2.2 to SAP BW 7.01. Thanks in advance pxbaker (BOB member since 2007-01-09)

120×120 10.4 KB

Got a requirement to change the names of the trace and error file names, which is possible using the -l and -z switches in the job launcher. When doing this they seem to by default end up in the dataservices/bin directory, rather than the ususal place. Is there an environment variable that controls this or some other switch? -l : Name of the trace log file -z : Name of the error log file (only if any error occurs) Also what does the -c switch do? Is this so that you can (for some reason) us…

Which is correct?
thanks

---

dnewton (BOB member since 2004-01-30)

[system]

Not sure which one suits best.
But I recently had an issue in 12.1 version with SAP Sources.

I had not touched the DSConfig. But I re-configured my Source and Target Datastores and it worked fine. (Werner’s )

---

ganeshxp (BOB member since 2008-07-17)

[system]

It’s version dependent.

Prior to 12.0 (11.7 and so on) it was your responsibility to pick a proper codepage for the engine. Default had the meaning of “use the codepage of the OS” which in my case would have been ISO Latin-1 (Western Europe + US). So I would have lost all the Chinese chars.

With 12.0 onwards default has the meaning of “engine should pick the proper codepage”. If all your source/target/lookup/transforms are ASCII only, why should there be a transcoder? Would cost performance only, the engine will run with ASCII codepage as well. When your source is Cyrillic and your target is French, neither the one or the other codepage can be used as each has unique characters the other does not support. So the engine will need to find the next best codepage which is a superset of both codepages - it will use Unicode.
And then some DQ transforms require Unicode, then the engine will be switched to the next best superset character set and for Unicode that’s Unicode.

---

Werner Daehn (BOB member since 2004-12-17)