> I think you are right.
> currently, the result is a series of child elements and the parent
> context information is not included in the result (except the namespace
> binding information that can be extracted into the split child
> elements).
>
> When I pushed the current implementation, I had the code sitting around
> in my repo to add such options of either wrapping the child with the
> its parent container part or set the parent container part in a
> separate header. But I didn't push that part, as I was not sure about
> whether this was useful (I mean, to be included directly in the
> splitter, as opposed to have this parent extraction step separately)
> and how to expose the configuration parameter for it.
>
> Probably including the extraction in the splitter is a good idea. I'll
> look at it again.
> regards, aki
I appreciate such an option.
We also have such a data structure containing of a root, a header and multiple
data nodes.
Jan
<root>
<header>
<...>
</header>
<body>
<data> one </data>
<data> two </data>
</body>
</root>
==>
<root>
<header>
<...>
</header>
<body>
<data> one </data>
</body>
</root>
<root>
<header>
<...>
</header>
<body>
<data> two </data>
</body>
</root>
