Hi,
This is the same structure used by QFJ when converting FIX messages into XML
using ToXML() method. So we simply used the same structure. See my comments
inline.
On Tue, Jul 1, 2008 at 3:51 PM, Asanka Abeysinghe <[EMAIL PROTECTED]> wrote:
> Hi,
> I have few questions about the FIX message structure maintain inside
> Synapse-1.2.
> Current message structure:
> *<message inSession="FIX.4.0:SYNAPSE->BANZAI" counter="1">
> <header>
> <field id="8"><![CDATA[FIX.4.0]]></field>
> <field id="9"><![CDATA[104]]></field>
> ..............................................
> </header>
> <body>
> <field id="11"><![CDATA[1213618394140]]></field>
> <field id="55"><![CDATA[SUNW]]></field>
> ................................................
> </body>
> <trailer>
> <field id="10"><![CDATA[226]]></field>
> ........................................
> </trailer>
> </message>*
>
>
> [1] What is the reason to divide the message in to three sections (header,
> body, trailer) ? Is it done based on looking at the FIX message structure ?
>
Yes. QFJ always associates a field with one of three message sections
(header, body and trailer). For an example to get a field in the message
header we should do something like;
message.getHeader().getField(fieldID);
The association is fully based on the FIX specs. So we are sticking to the
same convention. That is we preserve the logical structure of the FIX
messages when converting them into XML.
>
> [2] Why it has used repeating elements (field) rather than using unique tag
> id ?
> [3] Why it has used CDATA to store value ? (only tag 58 (text) might
> contain text we have to omit from the XML parser, not a common case)
>
Under most situations QFJ will allow putting any kind of data in any field.
Since Synapse is simply a mediator it should forward even this kind of
messages. (properly validating them and rejecting them should be done at the
FIX application that ultimately processes the message)
Thanks
Best Regards,
Hiranya
>
> The current structure is concerning me because
> [1] XSLT transformation and CBR make a tough task with the current
> structure, so it will effect the performance of FIX transformation and CBR.
> [2] FIX contains messages with repeating groups (E.g. Order List 35=E), IMO
> current structure will not support the repeating groups.
>
> Asanka A.
> --------------------------------------------------------------------- To
> unsubscribe, e-mail: [EMAIL PROTECTED] For additional
> commands, e-mail: [EMAIL PROTECTED]
