Re: V5R4 native xml processing - a user's critique

Hi Barbara,
 
<snip>
Yes. I'm not aware of any limit on the nesting level of the XML - I tried up
to about 20 nesting levels in my testing.
</snip>
 
That's great news! This means we can consume such things as SOAP messages in
one fell swoop. Again, nice work. :-)
 
<snip>No, it would pass it all as a single event. The data for an event can
be longer than 64K. (You'd have to handle it some other way than with an RPG
based variable in that case.)
</snip>
 
OK, that seems logical. If we have to handle single-element data over 64k we
can always split it across multiple elements. 64K does seem like a large
amount of data to place in a single element, but I have seen some
implementations that pass email data in XML and the whole body is held in a
single element. But, no problem. We can work with this model easily.
 
<snip>
Thanks Larry. I'm glad the documentation was clear enough for you to get
such a good understanding so soon.
</snip>
 
My pleasure - credit where it is due. Of course I'm just relieved we haven't
got a new XML() f-spec keyword and f-spec IFS manipulation. :-)
 
BTW, I think the ability to use xml-into with a %handler and a data
structure array will give RPG one of the easiest xml decomposition solutions
around. It will be so easy to simply specify the IFS document, internal xml
path (pseudo XPath), a data structure array (mapped (likerec) to an external
file format) and then simply code a for loop in the handler to write the
decomposed data straight from the populated array to the database table.
Fantastic! I can see xml as a nice container for transferring database data
between systems - and it should reduce the network bandwidth doing it this
way too.
 
One last question - the path option does not begin with a '/' in the
examples in the documentaion. Could I still write it with a leading slash? I
mean is the option 'path=root/level1/level2/child' the same as the option
'path=/root/level1/level2/child', or would I get an error if I coded the
last example? I'm used to XPath notation when navigating around an xml doc
and the leading slash jumps out as the document root and tells me that the
path is absolute, not relative. Can I do the same when coding the path
option, or do just assume that all paths are considered absolute and the
leading slash in redundant? 
 
Cheers
 
Larry Ducie