DSXchange

Note, in particular, the non-appearance of IN in that list.

You can change from fatal to warning in the Connector stage itself, for size mismatch and data type mismatch. They are two separate properties.

Oh, yes you are, which is why the error message is occurring.

For example what data type is this field in the reference data set?

Which repository are you asking about? The localuv data source is the project itself, the table is a UniVerse table (created with a UniVerse stage). Why do you need to find this object? Has anyone bothered to import the table definition for this table? If not, it probably does not exist in the repos...

A hash file is a tool for shaping a block of hash. In DataStage we talk (or should talk) about hashed files. The names are fairly self-explanatory. A static hashed file is one in which the number of groups does not change without manual intervention. In a dynamic hashed file the number of groups is ...

If you mean the Template (which can not be used as a project) restore from backup.

Upgrade to version 8 and you can use single login. Until then, not. You have to manage the users as operating system users in version 7.

At least two approaches suggest themselves.

1. Have another stream that bypasses the EndLoop activity. (This probably throws a warning "entered loop but did not exit".)

2. Set a flag that allows the in-loop activities to be bypassed if set.

There's actually no such thing as a default project. A project must be created when DataStage engine is installed, but there's no necessity to keep it once another project has been created. The rule is that you must have at least one project. If you feel you need it, create it afresh and import from...

At what point is this conversion taking place and are you expecting it to be performed implicitly or are you using an explicit conversion function?

If you are on a version earlier than 8.5 it matters hugely to the Transformer stage - all functions in the Transformer stage other than the null-handling functions are intolerant of null.

You can delete the hashed file (note: it's not "hash" file) using an after-stage or after-job subroutine. The actual command to use will depend on whether you are using an account-based or directory-based ("pathed") hashed file.

It's all about the speed of arrival of change events. Google can be your friend here.

Defined as not nullable on the reference input to the Lookup stage?

Agreed, that shouldn't happen. Have you reported it through your support provider? Well done on resolving it.