DSXchange

Seems like there are two problems here. One seems to be the fact that the lookup is failing. The hashed file is being read, however every row fails the lookup due to key mis-matches so your reference link total is zero after the job finishes. You are going to have to do all of the 'normal' things pe...

But still I don't know how database is accepting different rows with same ID Direct path loads are fast because indexes and such are disabled during the load, so it's up to you to not load 'duplicates'. The database will accept them just fine, it's when it tries to re-enable the PK constraint (for ...

If you've got a true fixed width file with no record terminators, I don't see how you can expect anything to be able to handle that in any kind of 'reusable' fashion. It needs to use the metadata to know how many bytes to read in for each record... and if that can differ on each file... Perhaps in P...

ppalka wrote:PS. Is this possible to create index in designer without using administrator?

Nope.

Actually, that syntax was specifically for hashed files under a UV stage. For a UV table, the other posts on the subject would have told you to use 'normal' create index syntax...

You could hint it to use a different index or use the NOINDEX hint:


to facilitate removal of the problem rows before you rebuild the index.

First thing would be to define what you mean by 'month'. Second would be to search the forum as this isn't the first time someone asked this question. For example, some answers here.

When I start my job without UV lookup it ends after 8min. When I start my job with UV lookup it ended after 15 hours <wow> Any ideas about what cause the problem? I've seen the same behaviour. UV lookups can be horribly slow out of the box if you drop it over a hashed file. As Kim noted, you need t...

Besides, we all know that the whole "the purpose of Dynamic RDBMS is to load into any database without changing the metadata, by just changing the database name, runtime" thing is a bunch of marketing whoey, right?

Not sure how much information you need, but in a nutshell: use the dsjob command. Most people will 'wrapper' it with a script to make it more robust, but technically all you need is that command. Read all about it in the Server Job Developer's Guide, there's a whole chapter on the Command Line Inter...

A Type 2 SCD should never need to UPDATE - you are always going to generate a new surrogate key value, and therefore you are always going to INSERT. Hmm... I'm going to have to beg to differ here. Typically, the Type 2 will include an effective date range and that is what is updated on the 'old' re...

The job has to abort in order for any uncommitted rows to be rolled back. And with a Transaction Size of zero, then 'any uncommitted rows' = 'all rows'.

Yes,elavenil ! You are right Not sure how that can be right when you made this statement in your very first post: I want to rollback all changes are made to databse if job returning with Status Failed from sequence. Transaction Grouping is used to roll back a single transaction when any of the link...

Transaction Grouping won't help for this as this is not the situation it was meant to be used in. Read the help on the option. Recognize that it forces the Transaction Size to be 1 for it to work. Then you'll see why.

You got the first two because of your code - the first is your message and the second is a direct cause of calling DSLogFatal. But I'm sure you knew that. The third message tells you the problem. You do realize that your tar command will untar the archive into your current working directory , yes? W...