DSXchange

This is what I came up with. Its working piece of code. More error handling can be added. FUNCTION getMedian(myList) Equate RoutineName To 'getMedian' myArr = CONVERT(",", @FM, myList) ; * Convert List to array numItems = DCOUNT(myArr, @FM) ; * Get the number of ite...

For a check you need two equal signs. You have only one ("else if (CHD_STATUS_REASON_CODE = 89)")

Line the data in sorted format and get the number dead in the middle. If there are only two entries, take the mean.

Next time anyone says Impossible, pass them this story and post what happened next

You can certainly tailor your own function that calculates a median.

You need more space. As Ray puts it, supply and demand.

A simple search would have revealed the methodolgy. Group changes and key assingment has been discussed a number of times.

My direction was towards modularization. More like an advice on not to do it in a single job.

Hmm, interesting. Server is still linient. Even if these functions are not used, the datatypes can be altered by specifying different target datatypes.

kumar_s wrote:There are 'Conversion' available under 'Functions'. Iconv/Oconv are majorly used.

ICONV/OCONV for datatype conversions , you sure Kumar?

Welcome Aboard, Batch jobs, in Datastage, is a collection of DataStage jobs coupled into a batch job. You can also write your own for more functionlity. Changing a batch job into a single datastage jobs means doing everything a batch job does in a single job. Not advisable, not recommened. Keep the ...

In server jobs just change the sql type of the target stage. The change is implicit, unlike parallel jobs.

The buildop will be c compiled code. So is the transformer's code. Ask them to give support a call for a document. The transformer has been imporved a lot and has no performance bottle necks. Go for a bench mark.

...and again.

Guys,
I recieved an interesting article/joke/story in my mail. It has a good moral to it so I decided to post it here. Check it out.