DataSet serialization

[AndrewD]

I'm curious if anyone here has used protobuf to serialize DataSets? I
saw an earlier discussion, but it didn't seem to go anywhere.
If DataSets are not currently serializable, how would I go about
making them serializable? My main goal here is performance and space,
not necessarily interoperability. I don't care if the DataSet doesn't
get serialized using a schema of some sort... we could just serialize
the DataSet as an object (for example, a DataSet contains a list of
DataTable objects, a DataTable contains a list of DataRow objects, a
DataRow object contains an array of object values, etc).
Is it possible in protobuf to define a way to serialize a 3rd party
object (one for which I cannot add the protobuf attributes)?

Thanks!
- Andrew

[Marc Gravell]

If you mean . Net data sets, I blogged about this earlier this week:

marcgravell.blogspot.com

Note I only experimented with a single DataTable, but that is most of the hurdle. The code using protobuf-net v2 is **experimental** and incomplete, but committed. If if is something you'd be interested in me pursuing further let ne know.

Marc

> --
> You received this message because you are subscribed to the Google Groups "Protocol Buffers" group.
> To post to this group, send email to prot...@googlegroups.com.
> To unsubscribe from this group, send email to protobuf+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/protobuf?hl=en.
>

[AndrewD]

Yes, I do mean .Net DataSets, and I am interested in this. Your blog
post is very awesomely cool to me! We will be at a point in a couple
of weeks were I will want to evaluate this for some alpha development
we are doing. If it isn't yet able to serialize a full DataSet, I can
certainly try to finish the coding for that.

Thanks!
- Andrew

On Oct 15, 2:14 pm, Marc Gravell <marc.grav...@gmail.com> wrote:
> If you mean . Net data sets, I blogged about this earlier this week:
>
> marcgravell.blogspot.com
>
> Note I only experimented with a single DataTable, but that is most of the hurdle. The code using protobuf-net v2 is **experimental** and incomplete, but committed. If if is something you'd be interested in me pursuing further let ne know.
>
> Marc
>

[Marc Gravell]

Note that I only includes enough code to handle the specific example in the blog experiment.

For the sake of the group, this discussion centres around an implicit contract inferred from a table structure, but persisted as a valid protobuf stream (a .proto could also be generated in theory).

Marc