DSXchange

What about the situation where your average input is a daily delta of 100K rows, your target table is 100M rows? Would you pull all 100M rows into a Merge stage just to throw away 99.9% of the rows to determine what's an insert and what's an update? Would it make more sense to first scan the 100K r...

The sort stage expects to do a numeric sort if a column has a data type of numeric and is one of the sort columns. Tell me, what number do you want A1 changed into?

You have a math expression somewhere that is not allowing for character data.

Using any of these comes to mind: CHANGE, EREPLACE, OCONV with the MCP token.

It could be that the NVL function is being called with ValeurDefaut not specified, meaning NVL(justoneargument) is being used. Or, the value of ValeurDefaut was never assigned when called, which causes errors like this. Your function looks valid.

Instead of directly executing dsjob use a shell script that runs the dsjob command that parses the dsjob exit code and returns whatever exit codes you want.

Welcome aboard! Functions return values, routines do something. A routine has an exit code to let the calling process know the status of the routine. A function is just a fancy calculation or derivation. This is function usage: x = myfunction(y) This is routine usage: CALL GoDoSomething Retu...

Depends.... If you have a table that is used as a reference in many different sets of transformations, do you want each job to pull that same table from the database every time or would you prefer a predecessor job pull it into a lookup dataset? What about the situation where your average input is a...

From your manuals under the Start button: You can use the Data Set stage: The Data Set stage is a file stage. It allows you to read data from or write data to a data set. The stage can have a single input link or a single output link. It can be configured to execute in parallel or sequential mode. W...

If you have two data sets that either can be the driver, but the goal is the INTERSECTION of the two (like an INNER join), then consider these scenarios: #1 Set A has fewer columns (or average characters per row) than B but more rows than B You want to use the A set as the reference and the B datase...

There is no Ascential supplied ODBC drivers for MS-Access on the Unix platform. You must either gain third-party drivers, use an indirect-bridge via something like SQL-Server, or use a custom set of drivers and write some data extraction scripts.

You'll have to make sure that your Unix box has all database OCI softwares installed with the correct library versions (ex: Oracle 32bit not 64bit) You'll also need to make sure that the jobs have all appropriate file types selected. When you designed jobs with Sequential file stages on the Windoze ...

If you build a reference lookup set of data that multiple jobs could share, write it as a persistent dataset (.ds) so that multiple jobs can leverage it. Otherwise, all jobs that need to access that data must also have all the necessary logic to go gain the data, and then redundantly use resources t...

Welcome aboard. Shame on Ascential for shoddy documentation of locks and engine commands. That's only been a common complaint since version 1.0. This forum is a great resource, of course, you could peek over at devnet if you wanted to as well. Stay awhile and stroll thru the oodles of posts regardin...

Link the MS-Access to SQL-Server, then use SQL-Server ODBC driver on your Unix machine to access SQL-Server. Or, consider finding the perl libraries that allow you to read .xls and .mdb files on Unix and use a script to pull out your data.