DSXchange

I am getting some very strange behaviour when using the field function with char(254) as the separator and I wondered if anyone could shed some light on the problem.

Here is the data:

Col1
Debt Consolidation~~2~448

Char(254) is represented by the tilde character for clarity

If I use the following expression field(col1, char(254), 1) I get "Debt" returned instead of "Debt Consolidation".

If I use the following expression field(col1, char(32), 1) I also get "Debt" returned which I would expect.

When I output column positions 5 (space after "Debt") and 19 (tilde after "Consolidation) I get char(32) and char(254) as expected.

So why is the field function using char(254) behaving as if it's using char(32)?


Look forward to hearing from you.


Many thanks,
Bob.

Just out of curiosity, what is the behaviour when you use system variable @FM rather than Char(254)?

Hi Ray,

I tried that, but the parallel transformer doesn't recognise @FM.

We also have NLS turned on and the following dsenv settings:

LANG=en_US.UTF-8
LC_ALL=en_US.UTF-8
NLS_LANG=AMERICAN_AMERICA.AL32UTF8

I have also viewed the data in hex and it appears exactly as I have described so I can't see that NLS is causing the problem.


Thanks,
Bob

This should shed some light on the problem...

This has to do with the Char() function. It must have originally been coded to handle only 7-bit ASCII characters.

When you give it a value outside that range, it defaults to using a space rather than aborting or warning.

Try using what the product documentation refers to as the allow8bits argument for 8-bit ASCII characters.

It is not very well documented as to exactly how to use that optional argument. It is expecting an expression that evaluates to a true or false value.