How to export/import multiple data base tables with relations to/from an xml file ?

Question

I have the following situation.

I'm building my application on top of a framework. I can access data only through web-services.

But the framework only contains web-services that return one table per web-service. In the data-base these tables have relation between them. For export (getting the data out of the database) I could easily merge the datasets returned by each web-service and then write the dataset contents to an XML file. How could I do this, but to take in account relations ?

It also have a set o web-services that can update the database content. I should be able to import the data I previously exported into the database. How could I do this ? (I don't think I can if I am exporting by merging the datasets).

What would be the best approach :) ?

P.S: I can't change the web-service framework. :(

10x

Answer 1

In my opinion, the best way to do this would be to add this functionality to the web services and not try to bolt it on from the outside.

---

I did have a situation like this once, where the web services in question were the salesforce.com API. In this case, I did build a DataSet with relations between the tables.

The problems were that if you enforced the constraints while loading data, then you have to load the data in the right order, and it will be slow. If you leave the constraints off while loading, and turn them on at the end, you'll get an exception at that time, but it can be difficult or impossible to identify the row causing the constraint violation.

Answer 2

I dont think you are going to be able to avoid writing some kind of a dal ... which means you need the tables as an object structure ... you might be able to use linq2sql for that?

then you could either use linq2xml or xslt to reconstruct the object structure and use linq2sql to write it to the db?

I know this isn't the easy way out you wanted, but i don't see how you can infer from multiple single sources with no necessary connection between them into a fully relational data model.

would love to see a better answer than this though, it seems like one of those holy grail questions.