DSXchange

Hey, no problem. I was thinking that maybe you had a local db object in your schema, but the full qualification was sending the query elsewhere.

Any chance your schema name is wrong? I created your tables and verified your SQL. The schema name messed with me at first (CNADW) but it works when removed.

Please post the exact SQL, as well as a describe of the tables involved. Please also use the code formatting so that it's easy to read. I'll bet dollars to donuts you have a column misnamed or mis-aliased.

Your reference lookup query looks syntactically correct. However, you did not paste the SQL that is giving you the problem. Could the missing GROUP-BY clause you left out have the issue?

There error message indicates that you are referencing a column that does not exist in one of the tables.

Read your manual under the Start menu. Specifically look at the Command Line section. The command line executable program is called dsjob. You can search this forum, it has been covered ad nauseum.

Assuming sysdate is returning a value of YYYY-MM-DD HH24:MI:SS, you have spaces, colons, and dashes that will not work. You must surround the entire string with double quotes. However, dsjob is sometimes flakey and will strip off the quotes. I have found that you have to use "eval" to execute dsjob ...

You must have discrete DS installations in order to have both domain installations running simultaneously. You cannot have two "engines" running jobs against the same underlying files. You can install DS on any domain in an E-series server, and each installation should have its own projects in its o...

sqlplus - run SQL statements

sqlldr - load data

I said I'd use a smart key for the surrogate, but that does not eliminate the need for a TIME subject area of tables. You still have to do a join to the appropriate table (D_DAY, D_HOUR, etc) to get to the attributes for aggregation/reporting purposes. I only use a smart key so that a date column is...

If you ask me my opinion, and keep in mind I've probably done ~30 installations/upgrades since version 3.1 on all available OSs, your best path is always a clean install. It leaves nothing lurking in the shadows or hiding in the corners/crevices.

Search the forum for uvregen. After changes to uvconfig, you must execute uvregen. This is covered alot on the forum.

The ORABULK stage is a wrapper to sqlldr. If you understand what sqlldr is, then you should know that it is not a data extractor, it is a data loader. Therefore, it has no ability to spool data, it is what is it is. Thus, you should not put an output link to stream data from that stage, as it not a ...

I can only point out that apples compared to apples should be identical. Imagine you have an empty non-partitioned table with locally managed indexes. You also have a 67 million row sequential file to load. If you use sqlldr with direct path option to load the file, it will read the file natively an...

The Time dimensions are the only family of tables where I would use a smart surrogate key. Simply take the dashes out of YYYY-MM-DD and you end up with a number. Take out the dashes, spaces, and colons from YYYY-MM-DD HH24:MI:SS and you end up with a 14 digit number. This is grey area, but I would a...