DSXchange

How do you define unprintable? Is it characters with ASCII values <32 and greater than 126? Any exceptions? How you code this depends upon what you need done.

Ok, you now have the 3rd option - neither NULL nor SQL.NULL but just an empty field. This will come into a VarChar field as '', or an empty string {which is very different from null!}. What Minhajuddin wrote now applies, but only sort-of. Basically, unless you are loading to a VarChar2 Oracle these ...

There is no single function in DataStage that will do this for you. It isn't a complex thing to do in either a Server BASIC program or a C++ routine, though. I would assume that the space character is the token delimiter and then write a loop to check each token with all following ones to check for ...

Ok, let us take things step by step. Normally this type of problem starts with mistaking an ASCII NUL and SQL NULL values for the same thing. Normally, a sequential file doesn't contain SQL NULLs. So as the first step let us see what kind of values we are dealing with. Use your favorite editor to lo...

Minhajuddin - what you recommend means that an empty string is equated to SQL Null, which I believe is not the poster's intent. Also, strings containing 0x00 (the NUL value) would not correctly be converted to SQL NULL (0x280).

I use both default values (for constrained columns like dates, times & numerics) and a transform stage with derivations for strings.

Adding indices usually increases query performance but slows down insert/update operations.

If you remove the lookup operation what does your performance do?

I'm not at a UNIX box right now, but if you go into FTP and type "help" there will be some command there that will display verbose output, including the ports used for communication.

You can concurrently write to the same dataset from different jobs; but you would need to sort your data explicitly when reading from the dataset. You would need to use the "append" mode and each process will create it's own partfiles inside the dataset.

The Before and After job subroutines will only accept 1 parameter. But it is a simple matter to concatenate strings with separators into the one parameter and then parse out those elements in your routine. The FIELD() function is usually the simplest solution.

Well, I just did the same thing in a test program and it worked like a charm, except when I miss-spelled a column and then I got a "The modify operator has a binding for the non-existent output field "TESTTER"." I can't see what might be causing this error. If you rename your input/output columns do...

This sort of logic should be done in an active stage before you get to the passive DB2 write stage.

Actually, you should get errors when writing nulls to a sequential file without handlers; so this means that you probably don't have a real SQL NULL but a UNIX null (0x00)

Ahh, it looks like the PORT command is the culprit. Are you sure that you are using the correct ports for commands and data in the FTP command?

What sort of a "Macro"? There are many different types and definitions of macros. In DataStage the term is not really used, I've seen it used in relation to DataStage Transforms but somehow don't think that this is what you mean. Excel Macros cannot be used in DataStage.