|
General Q/A
< Previous Next >
Re: field ordering
Colin Cullen / ICAP <>
11 Apr 2008 1:37PM ET Ian,
Yes we follow the first field in the group requirement that you mentioned, as well as the other rules defined in the FIX Protocol Vol.1 spec regarding "Repeating Groups". The engine we are using does treat RG content flexibly with regard to field order as you have observed with other engines.
Thanks.
Regards.
Colin
> The Message Format section in volume 1 of the 5.0 spec contains "4.
> Fields within repeating data groups must be specified in the order that
> the fields are specified in the message definition within the FIX
> specification document. The NoXXX field where XXX is the field being
> counted specifies the number of repeating group instances that must
> immediately precede the repeating group contents."
>
> This has been in the fix spec since the 4.2 errata.
>
> It should noted that many fix implementations dont implement/force this
> and engines often have validation/rule settings which allow support of
> repeating group fields in the correct or a mixed order. It is however
> imperative that the first repeating group element is the correct
> mandatory tag stated within the protocol message spec.
>
> > I wonder if someone could enlighten me regarding field ordering
> > requirements within FIX messages? The FIX Protocol 5.0 docs (what I am
> > currently using) speaks to field order only in regard to the detailed
> > documentation for Repeating Groups (VOL.1), where it says little about
> > field ordering except for the fact that the first field in the
> > repeating group is always required (because it is used as the
> > repeating group delimiter).
> >
> > I have also found a number of documents that state a FIX message must
> > be of the form <header fields> <trailer fields> with BeginString-
> > 8, BodyLength-9, and MsgType-35, being specified as the first three
> > fields in the header section.
> >
> > Neither our FIX engine vendor nor our application depend on
> > field order.
> >
> > I think it makes sense that field order requirements would be minimal
> > given the name=value construction of the message protocol at the field
> > level. The field name is what is needed to parse the message
> > (including any Repeating Groups).
> >
> > What is definitive with regard to this issue, and where is it stated
> > in the specification? If there is nothing definitive then what is the
> > prevailing consensus out there in the FIX community. Thanks to all.
Re: field ordering
Colin Cullen / ICAP 11 Apr 2008 1:37PM ET
|
