MQSeries.net :: View topic - cwf message, references, default values ?

novice · Posted: Tue Jun 21, 2005 2:00 am Post subject:

I'm having problems, respectively I don't know how to define a message with cwf-format, that has one element with a default value that is referenced multiple times.

My situation looks the following way:

I have to transform a cwf-message A into another cwf-message B. Both
messages have fixed length string elements. The elements of the message B
are additionally delimited by the string '|'. I'm not using a tds-format for
the message B because i want/have to use the padding functionality of fixed
length elements.

Here an example of a transformation:
Message A -> Message B
string1string2string -> string1|string2|string3

In the definition of message B I've defined a global element named Del,
which
is the delimiter. Within the properties tab of this element i've defined
that this element has a fixed value ('|'). Then I defined the structure of
message B the following way:
LocalElement1
Del (Element Reference)
LocalElement2
Del (Element Reference)
LocalElement3
â€¦
â€¦

Within the compute node, where i do the transformation i use the following
esql statements:
Set OutputRoot.MRM.LocalElement1=InputRoot.MRM.LocalElement1;
Set OutputRoot.MRM.LocalElement2=InputRoot.MRM.LocalElement1;
Set OutputRoot.MRM.LocalElement3=InputRoot.MRM.LocalElement1;

I want that the delimiters are included automatically.

This results in an output message that looks like this one:
string1|string2string3

The second delimiter is missed. The reason is that in my message definition
I use a global element as delimiter that I reference multiple times.

A possible solution would be to define each delimiter. In my case this not
applicable because the message B consist of a lot of elements. This would
result in defining (elements -1) delimiters, each one with a different name.

Another solution would be to define one single delimiter, that can occur
multiple times. This solution is also not applicable because then the
delimiters are not included automatically, since their position in the message is not defined. This would result in writing for each delimiter an esql-statement, which is not nice.

A third way to do it, would be to transform the message to the cwf-format and afterwards transform it a second time to a third format (tds). With the first transformation cwf to cwf the padding characters are included and with the second transformation cwf to tds the delimiters are included. In this case issues about the performance arise. Furthermore I have to define two formats per message.

Does somebody have an idea how to solve this?

By the way I'm using wbimb 5.0.5