I’m new on the forum. I tried to search answer to my problem in the forum but I was not able to find it…So I decided to make my first post…thanks a lot in advance to who will try to help me!!
I have the following problem (I’m using DataIntegrator XI)
I’m importing into a SQL Server 2K a text file, containing cyrillic characters and defined as fixed-length
I set my flat file and the datastore containing the target table using the codepage suggested into the tecnical manaul (1251, isoxxx and also UTF-8)
but I’m not able to load properly the files, because for certain lines -containing cyrillic characters-, DI consider the column longer than what it is in reality.
if useful, note that if I set as nvarchar the target column in my db I can see the correct value on it. also if I try to import the txt file in Excel, using the same settings for length I used in the definition of the file in DI, the file is properly uploaded (and values shown)
Can you post the trace logs please? I want to see what transcoders are used.
And when you drill into the file inside the dataflow, what codepage is used there? SQL Server datastore codepage should be utf-8 I believe in such case.
thanks for your answer. Unfrotunateley now I do not have the access to the trace log (I’ll post it not later than monday morning).
Anyway, for what I remeber, at the beginning of the job the code page is set to the default of the job server (the one that I found in dsconfig file). then, when the job goes into the dataflow the code page is updated. but I did not checked so deeply, honestly.
replying to you’re last question, I don’t know how to see which code page is used when I watch the file from the DataFlow. The only thing I recognized is that, when I filter the file for an item from the Dataflow (my file is a list of items) that I know has a description (the field) having cyrillic characters I’m able to see that the field is not well read by DI, because a part of it is put in the next field of the file bu DI. that’s why I think that probably the problem is or in the original file or in the way DI reads that file, not into the database. but it’s just my idea…
finally, you’re right, I can see something written properly in the database just putting both txt and Datastore code page to utf-8. this does not solve my issue (because beign the field considered too long, the following fields are not prepoerly loaded) but it’s a little step forward.
Both datastores for PKSAP are set to cp1252, that codepage does not contain Cyrillic characters, does it?
What database are those datastores for?
Can you change the database session setting to a codepage supporting Cyrillic and align the datastore codepage then?
In the meantime, I’ve open a request of support to SAP-Business Objects and they told that they are able to reproduce my issue in the 11.7.3xx versions of DI (we are using 11.7.3.2), while in release 12 the behaviour seems ok…
so it seems like the problem is on our version od Data Integrator
(BOB member since 2004-12-17)
(BOB member since 2004-12-17)