DSXchange

My problem is that there can be dirty data in any numeric coulmn for any of the tables that we are loading from our Legacy system, no way to know ahead of time. I'm trying to set a trap that will reject any coulmn. <paranoia>If there's a chance it will be dirty you have to check.</paranoia> One of ...

It's impossible to tell without knowing what the data types are and what data you are actually delivering. You may be able to determine this by inspecting the warning messages in the log, or you can definitely see the data by tracing the Transformer stage. The reason the intermediate text file works...

Why does that work? After your hard disk crash, any hashed file that was open was left in an inconsistent state. For example, the configuration table for job number 79 (RT_CONFIG79) possibly was left with inconsistent header information. Copying and compiling a job creates a new job, with a new job...

What was the previous value of RecsPerChunk?

It was right when I posted but, yes, they have reorganized the site. Now you can get it here (the $45 is for a paper copy - the PDF remains free). You can not directly query the VOC file for key columns of hashed files. Metadata are not stored in the VOC file; only the location of the data portion a...

The three dots indicate that there is more message. If you make the field wider you may see more. However, I'm guessing that there are explicit line breaks in the message, which means that only the first line is displayed in Summary view. This may be a change in behaviour since the previous release;...

You are using 2 output links but the second output link does not have any output columns. Is the problem because of that? When the second output link is a reject link from a Sequential File stage you can not dictate the metadata; you get one column, the raw input record. Something similar is true i...

My thought was that this computational load could account for the perceived increase in CPU consumption, particularly if a Transformer stage is involved (you are also carrying the call overheads to the generated library). How to trace and prove that may be tricky. Alas, I'm at a server-only site cur...

Can you please post one of the messages that you're describing?
Ideally one of each kind, though I realise that might be unreasonable to ask.

Either way you are splitting the processing. But you need to be clever to ensure that each process reads only a subset of the rows. Using a Link Partitioner, you still have only one process reading the file. This will remain an I/O bottleneck, even though spreading the processing may result in some ...

Is there a function in datastage that returns the datype (sqltype) of a field value as interpreted by the source during extraction? Cheers, Shyam No. DataStage is driven by the metadata. Therefore metadata need to be constant at run time. The information could be "hacked" from the design informatio...

A search of the forum will reveal that the DB2 stage expects DataStage internal format. Try @TIME by itself.

Otherwise, use the current time constant in DB2 SQL.

Try "MD" here means "masked decimal, and 2 is the number of implied decimal places in the input.

The next thing to ask is what is the setting in uvconfig of the parameter DISKCACHE?

Research the topic "Using Link Private Caching" in the Hash Stage Disk Caching manual (dsdskche.pdf page 10 particularly).

But I do believe you have grounds for complaint to the vendor.

Welcome aboard! :D Read more about link variables; you may find it more appropriate to use LinkName1.REJECTED to determine whether a row was rejected by whatever the Put() function is on output link LinkName1. But, to reject non-numeric input fields, an easier constraint expression would be Not(...