General

Re: Updating XML data

glebur — Tue, 04 Oct 2005 09:24:13 GMT

Hi Patrick and thanks for your answer.

quote:
Originally posted by Patrick Halstead

You will have to create some mechanism to identify which images, etc. came from the XML file so that subsequent edits can overwrite them. In a database, this is easy to do using a key.

I do have an identifier for each article in my database, and I use it as a prefix to the name of each file related to the article.

quote:
If you are shredding your XML to a database then uni-directionality should not be a problem unless someone else edits the data using a different client.

Only one person (the author) is allowed to modify an article, so that shouldn't be a problem.

quote:
Otherwise, it's an issue with your decomposition of the XML. In order to manage the update, approve, and re-publish, you need to create some kind of an index key or map to the data that was extracted.

User roles... sure this is easy to do and a common pattern.

In general, loops are easily supported even in Forms-based workflow. The issues are locking and data integrity if you have some line-of-business process that deserializes the data.

Okay, so you seem to be talking about a separate InfoPath form used for modifying the data, right ?
But how can I practically set up such a system ? I guess I have to create a web service method that works just like the publication method but the other way round, that is to say, it populates the modification form with the existing article data ?

If anyone has faced the same situation, it'd be great to have some experience feedback, too.

Re: Updating XML data

Patrick Halstead — Tue, 04 Oct 2005 06:09:01 GMT

Hi Glebur,
Great question and very eloquently stated. You will have to create some mechanism to identify which images, etc. came from the XML file so that subsequent edits can overwrite them. In a database, this is easy to do using a key. If you are shredding your XML to a database then uni-directionality should not be a problem unless someone else edits the data using a different client. Otherwise, it's an issue with your decomposition of the XML. In order to manage the update, approve, and re-publish, you need to create some kind of an index key or map to the data that was extracted.

User roles... sure this is easy to do and a common pattern.

In general, loops are easily supported even in Forms-based workflow. The issues are locking and data integrity if you have some line-of-business process that deserializes the data.

Sorry I couldn't be of more assistance,

Patrick Halstead [InfoPath MVP]
InfoPathDev

Updating XML data

glebur — Mon, 03 Oct 2005 12:21:55 GMT

Hi,

In all the Infopath samples and tutorials I've seen on the web, the workflow is limited to basic roles such as requester / approver, and once the contents is validated and saved, God knows what becomes of it.
In other words, the workflow chain is a uni-directional dead end.

But what if, once the XML data is stored, the requester wants to modify it over again (which means introducing a loop in the workflow chain) ?
Does he just open the XML data file ? Or does he have to use another form designed specifically for modifying the data ?

The problem is maybe a little more complicated in my case. I use an Infopath form to publish articles. When the approver validates the form, the data is submitted to a Web service that performs a number of operations (updating a database, decoding and saving the images...) and then stores the data minus the encoded images into an xml file on the server. That xml is later processed through an xslt transformation to display the article.

So how can a user update and modify an article using Infopath ? By directly opening the xml file stored on the server (maybe with an "update" tag) ? By using a different form that performs the reverse operation, putting all the data and images back into a single InfoPath data file ?
And most of all, how can Infopath help my web service detect that the article is a previously published one and that it has to write over the old files rather than creating new ones ?

Besides, is it possible to integrate this in the user roles, for example defining that when a user modifies an article, that user takes the requester role and the contents has to be validated again by the approver ?

Thanks for your help...