|
why not use XSLT?
"What classes are you using ? You shouldn't call stuff if you have no idea what it does"
Christian Graus in the C# forum
led mike
|
|
|
|
|
led mike,
Can you explain how I would use XSLT to generate these XML documents from empty example ones? Do you mean to transform the empty document into one that is populated with the data values?
More background: the data that I will use to populate the values in these XML documents will be coming from an ERP system (ODBC).
Thanks for the recommendation and clarification.
Josh Blair
Evergreen, CO
|
|
|
|
|
Sorry I misunderstood the first post. It sounds like you have 2 issues.
1) Generate XML data from ODBC data source.
2) Not have hard-coded XML format (element and attribute names)
First, there must be a fairly wide variety of options to accomplish those goals. I certainly do not know all of them, here are a few I am aware of.
Before I start in on a software development solution, you might look into Biz Talk Server. I have never used Biz Talk myself (only read about it) but it is highly possible it would provide a completely configuration based solution.
There may be other products as well however if for whatever reason you need to develop a solution here are some ideas.
I believe #1 is simple because I think there is a Microsoft ODBC driver that is compatible with ADO.NET. Look on MSDN for that. Obviously it is not the fastest driver.
So if we assume you can use ADO.NET to access the data then you can use the XmlSerializer to generate the XML. The bare bones generation will use the table and column names from the data source to generate the XML. Assuming that works for you then you are done and nothing is hard-coded.
If those names do not meet your requirements then you have many choices you can use to change the element and attribute names. There are Attribute (Declarative) mechanisms you can use to effect the Serialization names that are created. Again you can read about this on MSDN.
In another potential solution you could use the raw XmlSerializer XML and process those results into the desired XML using XSLT. Then the XSLT becomes in effect your "template" that can be changed to modify the end result XML without any compilation required. However it would be somewhat slower than another solution, Custom Serialization.
Basic custom serialization would in effect be hard coded. If you wanted to change the output XML format it would require compiling. However you could develop a system that reads in an XML format or mapping document and uses that to generate the output XML. This is more work but would result in you being able to change the XML mapping documents to change the output XML without any compilation involved.
Hope that helps
"What classes are you using ? You shouldn't call stuff if you have no idea what it does"
Christian Graus in the C# forum
led mike
|
|
|
|
|
led mike,
Thanks again for your information. It is very helpful.
I must not be stating the problem and desired results appropriately. I apologize. I will try to explain myself better...
Good news: I wrote the code generator that I was looking for myself. It is a simple solution that I may post if anyone is interested. I'll give you a description of what it does: It is a console app written in C# that takes a path/filename to an example XML file as a parameter and then subsequently generates the code necessary to create a duplicate xml structure in code using the XmlTextWriter. Very simple but effective. Once this generator does its thing, you can put the output into your own class files and use them as a starting point. You can add public properties and the like to these class files and then you use them to generate XML files from ODBC queries or any other datasources. Does this make any sense? I still my not be articulating this very well.
Here is an example of what you can do with these objects once you get them built (pseudo code only):
XMLInvoice invoice = new XMLInvoice();
invoice.InvoiceNumber = "INV12355"; // may come from helper method or database query or something like that...
invoice.OtherField1 = "whatever";
invoice.OtherField2 = "whatever";
. . .
// in a looping construct that may result from records obtained from a query or the like
invoice.Lines.Add(param1, param2, etc...);
invoice.Lines.Add(param1, param2, etc...);
invoice.Lines.Add(param1, param2, etc...);
. . .
invoice.CreateXML(); // this is where the XmlTextWriter is used to generate the big nasty XML message (that the code generator helped create)
. . .
invoice.Save(filename); // for archiving purposes
. . .
B2BEngine b2b = new B2Bengine(); // object that knows how to communicate via HTTP with customers, suppliers, etc.
CustomerInfo custInfo = new CustomerInfo("MyCustomer123"); some settings info about particular customer (Ip address of server, credentials, etc.)
b2b.Send(invoice, custInfo); //
Please me know if this helped to explain my situation an better. Also, this may lead you to believe that I am approaching the problem all wrong. If that is true, please let me know your thoughts...
Thanks again,
Josh Blair
Evergreen, CO
|
|
|
|
|
Josh Blair wrote: I must not be stating the problem and desired results appropriately.
Yes I still don't get it. Based on your latest post I still think your Goal is:
1) Generate XML data from ODBC data source
Is that correct?
If it is correct then yes I do believe it is possible that you are not taking advantage of the .NET Frameworks inherent power to accomplish this. The Library has very rich capacity to connect and consume Database data. Just as powerful are the XML libraries such as the XmlSerializer that I mentioned.
"What classes are you using ? You shouldn't call stuff if you have no idea what it does"
Christian Graus in the C# forum
led mike
|
|
|
|
|
led mike,
To answer #1, let me use an example: lets take an invoice...yes some of the data comes from an ODBC datasource (ERP system), but only some of the data. Other portions of the data come from other sources of data as well (config files, other data sources like SQL Server, etc.).
I've used serialization in a limited capacity for things like reading and saving application configuration settings and such.
Are you saying that I should build business objects that are serializable? As in an Invoice object that is serializable so I can instantiate the object, fill it's properties, call it's methods and thn have the ability to serialize it to produce an XML document from this business object?
Thanks again for the clarification,
Josh Blair
Evergreen, CO
|
|
|
|
|
Here is something that might help: different portions of the xml document(s) will come from different sources of data. Again taking an invoice as an example (but there are other types of docments as well):
I have to send invoices to a third party. This third party defined the xml structures that I have to populate and did not provide schemas or DTDs. A large portion of the data that I need to send to this thord party will originate from a legacy ERP system that does offer an ODBC driver. I can obtain the data from this system, no problem...already got that part licked.
The invoice XML document has a "header section" which contains sub-sections like "ship-to section", "buyer section" and so on nested within the "header section".
It also contain a "line item section" where there can be one to many line items in this section.
Some of the data that populates the "header section" will come from the invoice header in the ERP system (via ODBC). Other parts will come from a configuration file, while other data will come from database tables not originating from the ERP system (maybe a SQL Server).
Most of the data that will populate the "line item section" will come from the ERP system from the lines on the invoice.
Serialization might be the proper approach. I think the code generator that I wrote may be misusing the .NET framework because it is essentially reproducing the features of serialization. BUT...when using serialization, bdo you have full control over how things get serialized? I can't have all the "junk" that comes with serialization out of the box. When I say "junk", I mean the xml namespaces and extra crap not allowed by this third party. As in: xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance".
Also, I need to have full control over the order of how elements are ordered and nested. Is this possible with .NET serialization? I'd rather not have to transform the serialized business object before I send it off because I'm not really versed at XSLT.
Josh Blair
Evergreen, CO
|
|
|
|
|
It sounds to me like you could use a DataSet to coalesce the disparate sources. See some DataSet articles I know there are some on CodeProject. Josh Blair wrote:I need to have full control over the order of how elements are ordered and nested.
A DataSet fulfills that requirement. You can design the tables and relationships to your requirements.
Josh Blair wrote: BUT...when using serialization, bdo you have full control over how things get serialized?
Yes there are several different ways to control serialization.
Josh Blair wrote: When I say "junk", I mean the xml namespaces and extra crap not allowed by this third party. As in: xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance".
That makes no sense at all, those are XML standard namespaces.  Anyway that should be possible one way or another. I expect you can control that using declarative attributes in your code. Look at the XmlSerializer and the XmlRootAttribute attribute.
|
|
|
|
|
Josh Blair wrote: Other portions of the data come from other sources of data as well
ahhh... didn't know that.
Do these source data schemas change?
Why do you keep refering to XmlTextWriter? Do you have some requirement to use it? Why?
|
|
|
|
|
led mike,
Not sure if these XML structures (schemas) change very often. They come from a company from Japan (from a division in Italy). I think they are pretty solid/permanent. Also, I'm not sure why they didn't provide schemas or DTDs either...
I am using the XmlTextWriter only because it is fast (not that speed is really and issue at this point). I used the XMLDocument object in a previous cXML project that was very similar to this project and it seemed a little slower to generate the xml docs. I just wanted to try the XmlTextWriter to see if it was noticably faster. It is also easier to use the XmlTextWriter in my code generator. I also thought about the StringBuilder and it work just as well probably. The XmlTextWriter has a fwe handly features like being able to close off the element and stuff like that.
Josh Blair
Evergreen, CO
|
|
|
|
|
I think I understand what you're talking about, although it doesn't make much sense. You take an XML document as input and create the code that will generate that XML document. I've never seen anything like that before, but I'm pretty sure you could just write one yourself using an XmlReader and a StringBuilder (or better yet - CodeDom).
|
|
|
|
|
Yes, the output from the code generator alone doesn't offer much by itself. So then, once I have the code that produces the XML document (an empty version with no invoice data), I add to that code. I add public properties with get and set accessors. I add the ablity to add line items to a "line itme collection", things like that. So from my code, I can grab all the data necessary that will eventually populate this document, then I instantiate this xml invoice business object filling its properties, collections, etc. Then i call a "create" method which generates a new xml document with the proper structure containing all the invoice data from my ERP system and other data sources. Am I approaching this bass-ackwards?
Thanks,
Josh Blair
Evergreen, CO
|
|
|
|
|
|
Dustin, thanks for the link. That is a pretty cool app. I tried it with my XML structures and it fails. I tweaked the app and my XML a bit an got it to run. That is a great sample. I'll see if the output is going to be useful and let you know.
I also tried running XSD.exe (mentioned in one of the comments to the article) over my XML samples. Here is the output:
C:\Projects\Customer_XMLCodeGenerator_CS>xsd CustomerDespatchAdvice.xml
Microsoft (R) Xml Schemas/DataTypes support utility
[Microsoft (R) .NET Framework, Version 1.1.4322.573]
Copyright (C) Microsoft Corporation 1998-2002. All rights reserved.
Error: There was an error processing 'CustomerDespatchAdvice.xml'.
- The same table (Country) cannot be the child table in two nested relations.
(In this xml sample, there are multiple sections with the same element (Country), and the tool treats those as "tables" and 2 tables can't have the same name...)
I have run into this type of error before with other XML-based projects. I think what is basically going on is that the XSD tool determines that the XML data is not valid according to the ADO.NET schema inference rules. See the following article for more on this issue. Quite a limitation if you ask me. Not a big deal if you have the luxury of defining the XML data from the beginning but not cool if you are forced to use someone elses structure...
I had the same problem with cXML (http://www.cXML.org in another similar project)
http://support.microsoft.com/default.aspx?scid=kb;EN-US;325696
Anyway, thanks for the tip!
Josh Blair
Evergreen, CO
|
|
|
|
|
Glad that helped.
We had a similar problem at one of my previous jobs. We wanted to just take their XML format and dump it into a DataSet, but it didn't like the "duplicate" table names. And we couldn't change the format. They ended up creating custom classes that handled the serialization. The problem I had with that solution though was that there wasn't much code generation going on to create those custom classes and there were no unit tests to make sure they worked 100% correct. If there was ever a good place to use code generation...
|
|
|
|
|
Hi all,
I have a very weird problem:
I have a web page (myfeeder.asp) to generate very simple xml content. I use ie to navigate to the page, I can see the xml content on MOST of the pc. the xml like:
<?xml version="1.0"?>
<ads>
<ad>
content here.
</ad>
</ads>
My problem is, I CAN NOT get the xml content from some of web server's IE. what I got is nothing or just the structure, like below. It also happpends to some linux server.
<?xml version="1.0"?>
<ads>
<ad>
</ad>
</ads>
I know the server/pc that can not get the xml content is using re-writing.
I have no idea what cause the problem!!
Does any one have the same experience!!! Please help me to solve the problem!!!!
Thanks in advance!!
Alan Shen
MCAD for .NET Version
^~^~^~^~^~^~^~^~^~^~^
Great idea is the beginging of success!
-- modified at 23:04 Wednesday 24th May, 2006
|
|
|
|
|
Hi all first time caller ; long time listener;; haha ;Had to say that...
Here is my deal; i have worked with xml and schemas now on and off for about 16 months and i am just now really seeing the power of xsd and xml ....
I am modeling my class objects with the xsd designer in vs.net 2005 and then run the xsd.exe to produce my class obj;s in c#. When i am designing a complextype when needed i will create an unnamed simple type as an enum..
Here is my question? I need to design an xsd that will have complextype like.. MaterialType that has an element of blinds( here is where i need the help -->)that has a group OR enum of (WOOD, plastic, faux, ECT...); that has a group OR enum of ( 1/2", 1", 1/2", 2/1/2", ECT...) that has a group OR enum of (color, White, red, blue, stain( NOW even the colors may have a need for a group or enum like( LIGHT BLUE, ***** Prices can be base on color options that is why i need so many levels,, and (or) prices based on Wood or plactic..)))
I think everyone can see where i am going and i know this is a simple one for someone who has created a schema that will offer the most ease of development...
Can someone tell me if i need to work with the GROUP, ENUMS, ?????? Looking for any and all suggestion here?
Erik
|
|
|
|
|
I'd say that embedding things like colour, size etc as an enum is probably hard-coding too much - what happens when extra colours, sizes are added to the product range?
I would tend to use a relational database approach rather than multiple levels of hierarchy - define the range of colours, sizes etc separately and then link to them somehow (ids and id-refs, probably).
|
|
|
|
|
Thanks you for your help! Very much appreciated.
Erik
|
|
|
|
|
I did something similar where all my different enumerations were each a single simple string type - pen width, colors, line styles, etc. Then I made a complex object containing a sequence of elements where each element was one of the enumerated types, each type represented a single time.
People that start writing code immediately are programmers (or hackers), people that ask questions first are Software Engineers - Graham Shanks
|
|
|
|
|
I want to get some examples in ASP using XML that how connection has been created using XML in ASP / Javascript.
Please give me proper suggestions.
Thanx
Jesal Rana
OK
|
|
|
|
|
|
Please help me out to get a particuler value from XML document.
I want ti dril down in an XML doc to search for a value.
Urgent .
sanjeev
|
|
|
|
|
See XPath[^]
"What classes are you using ? You shouldn't call stuff if you have no idea what it does"
Christian Graus in the C# forum
led mike
|
|
|
|
|
if i have to manager hundreds (maybe even thousands)of xml files, now i wanna offer my user a search function, is efficiency a serious problem f i use SAX or DOM to search for information the user query for? should i open the file just in plain text mode?
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
