Back to Questions
5.1k
views3
comments
F# to parse XML using dataset.
f#webscheme

Hello, I have to use F# to parse items in XML format from internet.
To avoid the trouble, I downloaded the XML file and saved it locally and made it available for local web browser.
The URL is: [link:localhost]
The following is just one part of the file, the actual file is 100+ times as big as this one.
<?xml version="1.0"?>

<items>
<item partNum="872-AA">
Lawnmower
<quantity>1</quantity>
<USPrice>148.95</USPrice>
</item>
<item partNum="926-AA">
Baby Monitor
<quantity>1</quantity>
<USPrice>39.98</USPrice>
</item>
</items>

When I tried to use ordinary XPATH to query the file:
let xmlpurchaseOrder = GetWeb("[link:localhost]
let xmlOrders = new XmlDocument()
xmlOrders.Load(xmlpurchaseOrder)
let xmlItems = xmlOrders.SelectNodes("//purchaseOrder/items")
printfn "Done"
The GetWeb() function is just to read the contents of the web page into a string and return the string.
However, my code didn't work, I got the error message: Invalid URI: The Uri scheme is too long.
As I know the file contains 100+ items, so I can understand may be it is too big to do it this way.
When the XML file is short, my code works!

I found from internet that I can load the Xml into a StringReader and then read it into the DataSet for big XML files.
Here is my new code now:
let xmlpurchaseOrder = GetWeb("[link:localhost]
let ds = new DataSet()
let sr = new StringReader(xmlpurchaseOrder)
ds.ReadXml(sr) |> ignore
sr.Dispose |> ignore
printfn "Done"

The code worked.
But now I have another headache: how I can parse the purchase order items just as I used my old code did?
Thank you for your help!

.

let xmlpurchaseOrder = GetWeb("[link:localhost]
let xmlOrders = new XmlDocument()
xmlOrders.Load(xmlpurchaseOrder)
let xmlItems = xmlOrders.SelectNodes("//purchaseOrder/items")
printfn "Done"

use 

1

xmlOrders.LoadXml

instead of 

1

xmlOrders.Load

. The argument of the Load method is the path of the content not the content itself.

By on 3/6/2010 6:05 AM ()

Hello, Sharpdev:
Thank you very much.
Finally, someone has answered my question. Actually, one of my friends has already told me the issue by email. But it is difficult to find anyone in this forum to indicate this.
Thanks again!

By on 3/6/2010 7:32 AM ()

mainly because that is strictly speaking not a F# question and more about a .NET library thing.

That said, it highlight one of the nice thing about F# unit of measure(not sure that applies to string though) is that for generic data primitive type like int/float/string, there is a need for unit of measure so the compiler can do some type checking. that is string <> string<xml>. That though is more about function parameter design and not much can be done in this case anyway.

By on 3/6/2010 9:51 AM ()

    Topic tags

    IntelliFactory Offices Copyright (c) 2011-2012 IntelliFactory. All rights reserved.
    Home | Products | Consulting | Trainings | Blogs | Jobs | Contact Us | Terms of Use | Privacy Policy | Cookie Policy     		Built with WebSharper