DSXchange

Enable reject links, and output that to a file and include the SQL Error code to find out the cause.

There are only two basic methods of getting duplicates, either sorting the data by the column in question and comparing adjacent rows, or by keeping the whole contents of the file as it is being read in memory and doing a comparison (i.e. array or table scan). Either method may be quicker and more e...

That means that with VarChar(60) still only 6 characters out of 10 arrive in DataStage? Are there any NLS settings or special characters involved? Can you reproduce the problem with a new SQL Server table or just the one that you have so far?

There are various numbers floating around regarding Gb per CPU and I've found that the true production numbers depend greatly upon how the application has been coded, both for server and for parallel systems. There is no upper bound or cap on CPU usage, but again depending on your application and co...

I had that issue a while back and had to use the length and an IF THEN ELSE clause with two different pictures for the date. I vaguely recall having seen someone post a format that accepts variable length date components but am not sure - I do know that it isn't documented.

Enable reject records in the Oracle stage and then put an output link; from this stream you can do such things as determine the errors and also do a count and write that to another location.

Use a 2-stage process with a loop in your UNIX script,
a) "dsjob -ljobs MyProject" to get a list of jobs in your project
b) for each job, use "dsjob -run -jobstatus MyProject MyJob" to get the status.

Variable $R246 seems to be undefined. Could you search the output for that and see if you can figure out where it is assigned a value?

$R246 seem to be undefined. Could you search the output for that and see if you can figure out where it is assigned a value?

DataStage functions identically on both platforms. So the question is more of a "Windows vs. UNIX" one with all that that entails. That is almost like asking if one political party or the other is better...

Set the default value of the column to "0" in the column attributes on the sequential stage and the problem will also take care of itself. I personally prefer to use the suggestion of using a VarChar column and explicitly changing the value, but defaulting is quicker and simpler, just not ...

concurrent writing in append mode to a dataset works.

Ok, let us assume that your stages are all necessary. Then what you need to do is ensure that the steps happen sequentially instead of in parallel so that the scratch space doesn't fill up. Or that less processes use scratch. How many nodes does this job run on? What happens if you use a 1-node conf...

How many nodes are you running this on? Is your database partitioned? These factors can affect your runtime as well. The same applies to the order in which the sparse lookups are done as ordered queries might retrieve cached results.

Steve - that sounds like a reproduceable test case for IBM support. Do you have an Oracle instance floating around to test on; I wonder if it is DB2 specific or sparse-lookup specific.