DSXchange

1967-12-31 is day number 0.

There's no function you can use in a watch unless you - perhaps - add a dummy column in a Transformer stage and discard it in the following stage.

Even fix pack 4 is in the market.

Can you post the actual schema file? Can you please also post the score from one run of the job with more than 100 records? There may be a limiter there. What stage type are you using to read the CSV files?

Search DSXchange for source code for pxEreplace() - it does not ship with DataStage so you have to create a parallel routine of it.

Directly is OK for 7.1 to 8.5. It is not necessary to upgrade to 7.5 first. Equally it's not wrong to do so, if you want to.

Has the job actually been run in reset mode?

Check the parameter type (e.g. Date) then make sure a valid value for that type is being provided. Date must be in the format YYYY-MM-DD.

It is easily possible to do this with Ereplace() if that's the only occurrence of 301 in the data, otherwise you'd have to do some more manipulation to determine which occurrence of 301 to replace. Replacing the substring is likely to be a lot easier. InLink.TheString[1,9] : "801" : InLink...

Case by case basis. I guess anything under 1-2 MB (possibly larger) would fit in the small category. There's nothing to prevent a server job setting up a parallel load.

Not this year. I am fully committed through till at least November 2011 doing non-USA things. Probably beyond that.

Assuming that my assumptions about the arguments are correct, I would probably design the code to open the file, read each line in the file and output the fields mentioned in the second argument. FUNCTION Grabit(aSubjectValue, aFilePath, aKeyValue, aKeyPosition) $INCLUDE UNIVERSE.INCLUDE FILEINFO.H ...

Directly is OK. But you'll need rather more hardware capacity for 8.5 than you needed for 7.5: the extra is taken up with an extra database server (DB2, Oracle or SQL Server) for the common repository, and an instance of WebSphere Application Server to supply the services tier. With version 8.5 you ...

So it's not conversion, it's composition. Is this correct? What will be the arguments to this routine? I'd imagine them to be the pathname to the file, the subject value on which to filter, the key value on which to filter and a specification of which fields are to be returned in the form "1:3&...

Tough. You simply cannot look up NULL. It is not a known value. (That's what NULL signifies.)

Here's an analogy: I have something in my hand. Tell me what it is. (That's the equivalent of a lookup - you have no way of knowing what I have in my hand.)

You would be using the correct function if you actually went ahead and tried it!