DSXchange

Import as a single varchar record.

Filter/Transform based on the first set of characters.

Column Export/Transform to split up the columns in a specific format.

Play on.

Have you ensured that the log is not saying anything negative (such as dropping records)?

I believe there IS an option to export as XML. However, I do not know what this option are. I do not have access to a server at the moment to experiment with this.

Play around with that script chulett suggested. Hope you know DOS batch script language.

There is apparently a limit on tablename set somewhere. Thus, the tablename is invalid/truncated/whatever.

I do not know how to resolve this, other than to try shorter tablenames.

You have a non-nullable field that was provided NULL. Whoops. DataStage's default behavior is to drop the entire record.

Do a search for a recent posting about this. I ranted pretty good, and provided some tips on how to figure out where the errors really are.

(APT_DISABLE_COMBINATION)

Now put down that Universe weapon and slowly back away. It's dangerous to manipulate your own project tables directly, even if you think you know what you're doing.

This illustrates the problem with the GUI interface to the original Orchestrate engine. On Orchestrate, all you have to really do is define the input and the output, and only the fields you need to transform throughout the OSH code. However, the metadata does not show up on the GUI, so it is difficu...

It means what it says: The input dataset does not have "name". You may have defined it in your metadata, but if the database don't have that field, it won't show up.

Use the DataSet Management tool to observe the defined metadata for the input dataset.

Use the following command: $APT_ORCHHOME/bin/orchadmin delete [dataset header file] This will delete not only the header file, but all the reference files. You do can point those dataset files elsewhere using your APT_CONFIG_FILE configuration (Disk). See the file "man_gde.pdf", chapter 11...

If you want to count the months, you can get MonthFromDate(). If you want to count the days... MonthDayFromDate(). There are a number of options you can use -- just look at Appendix B ("Functions") on your Parjdev.pdf file.

nuff details?

There's many different ways to skin a cat...

parjdev.pdf - Appendix B - Functions.

Post here if none of them seems to fill your requirements, and we can present you with solutions.

However, first use a search to make sure nobody else already asked and we answered.

Oh great. Heh. I wish I have access to Oracle to help. Sorry.

All up to 7.5 does not support flexible naming convention. Just insert a copy stage before one of the links, and rename the field.

Arand is correct. The types you mentioned are generally stuff that are provided by default per installation. You did not have to extract EVERYTHING. You can just extract the job design if you are willing to recompile on the production code, or include the job executables. That's all you really need ...