DSXchange

Yes you can lookup to the same hash file multiple times in the same transformation. You just need to have three links from the hash file to the transformation so that you can define each link. You will also have to define the output three times in the hash stage for each of the output links.

We do this type of thing all the time. Dimensional models work great but sometimes the reporting requirements cause you to look for other options. You can have the greatest dimensional model possible but if it takes too long to generate a report, the users are not really going to care. Perception is...

Traditionally data warehouses are setup so that the warehouse controls pulling the data from the source systems. This is done so that the processes that build the warehouse are all controlled in a single place. In your example you have multiple source systems. You would not want to change each one o...

Your lookup to the product code hash file is not working. You said that when you build the hash you are trimming the product code. Make sure that the value you send to lookup the code is also trimmed. If there are any trailing spaces you are not going to get a match and everything will be reinserted...

You can use the pivot stage to split columns into rows. It has been covered here many times before. Just do a search on 'pivot' or 'horizontal split.'

If you really want to use the lookup you have the option to return multiple rows if you're looking up to an ODBC stage. In the Input properties of the transformation stage there is a check box titled 'Reference link with multi row result set.' Check this and your lookup will return multiple rows if ...

If that is the result you're get then something is wrong with your join. The results that you are displaying are accurate if you're reading from table1 and doing a lookup to table2. They are not accurate if you are joining the tables as I suggested. The join will pull all combinations from your two ...

Based on your example your just doing a simple join between two table. You can do that in any stage including OCI, ODBC, ect.

The additional fields that your seeing were not added by Datastage, they were added by someone who can create and modify tables in your database. One way to tell is that Ascential is a little more more consistent with their naming standards. :wink: I'm guessing that the physical table layout in your...

The problem is in your insert/update statement since in the error is happening during the Put. This could be caused by something as simple as trying to update a key that isn't there. Look at the log through Datastage director. It should give you more information including the exact record that is ca...

Stages like the ODBC stage do not have a delete option in the list of update actions. In order to delete you have select user defined sql and then write the sql statement yourself. Pass in the key column to your delete stage and code your statement similar to the following: Delete From "My_Tabl...

This sounds like an application where you could use Ascential RTI. You would get the benifit of using the business logic contained in your datastage jobs without being tied to a batch schedule. This of course assumes that you have an unlimited budget.

You need to create your ODBC DSN on the server running Datastage, not your individual client machine. The client for SQL server is the Datastage server.

The datastage CDs contain a tutorial that you can go through. I don't remember exactly where it is but they are not too hard to find. They basically go through creating a simple project using a database that they also provide. It is pretty basic stuff but it is what I used when I started. It provide...

Could you just bring the same file in ten different times and send those to the link collector? I'm not sure if that would improve performance but it would remove the process of creating ten different temporary files.