ansaurus

Question

Get a 2 String array from HTML .... using regex?

Answer 1

+2 A:

You should use System.xml and use a proper parser to do that work. Xpath or even navigating in the XmlDocument would permit you to achieve what you are looking for.

Dim xpathDoc As XPathDocument
Dim xmlNav As XPathNavigator

Dim xmlNI As XPathNodeIterator
xpathDoc = New XPathDocument("c:\builder.xml")
xmlNav = xpathDoc.CreateNavigator()
xmlNI = xmlNav.Select("//span[@class='mainTextbold']")
While (xmlNI.MoveNext())
    System.Console.WriteLine(xmlNI.Current.Name + " : " + xmlNI.Current.Value)
End While

I suggest you to take a look there or there for more info how to extract information from a XmlDocument

a Xpath selector like span[@class='mainTextbold'] would return you all those span.

as per Heinzi remark, your document doesn't seem to be valid XHTML, you should convert it to XHTML using TidyNet and then parse the result of the conversion.

RageZ 2009-11-20 06:48:22

Ok, I should have mentioned that this is being returned to me via VS 2008 built in WebBrowser module as an actual webpage.So I assume xpathDoc = New XPathDocument("c:\builder.xml") would become something like xpathDoc = New XPathDocument(webbrowser1) Sorry if I am way off here...I've been using functions such as WebBrowser1.Document.GetElementsByTagName("THETAG")andcurElement.GetAttribute("THEATTRIBUTE").ToStringTo help locate Textboxes, dropdownlists ect to auto fill the form.

Travis Walker 2009-11-20 07:03:02

going to dig through those links in the mean time and see where that leads me. I knew regex could probably get the done but I was certain there was a better way. now it seems I am at least on the right track. I'll check back here shortly.

Travis Walker 2009-11-20 07:05:04

The HTML posted in the question is not valid XML (see, e.g., the unclosed <BR> tag...)

Heinzi 2009-11-20 07:07:01

Answer 2

+1 A:

To answer the original question, taking into account all the mandatory caveats about "parsing" HTML with regexes, here's a regex that will do what you want:

Ref#: (.{9})[\S\s]*?Label Number: <span class="mainTextbold">(.{27})

Backreference \1 will contain the 9 characters after Ref#:, \2 will contain the 27 characters after Label number...

Alternatively, to make it a bit more robust, you could also use

Ref#: ([^<]+)[\S\s]*?Label Number: <span class="mainTextbold">([^<]+)

That way, the regex will match any characters except opening angle brackets after the markers. It will lead to a lot more backtracking in case of strings where the regex can't find a match at all. Depending on the regex engine used, this can be avoided if you use possessive matches:

Ref#: ([^<]++)[\S\s]*?Label Number: <span class="mainTextbold">([^<]++)

The rationale behind my support of using regexes for this task:

it's trivial and easy to read/maintain - arguably easier than parsing code
there's only one match per page, no nesting.
it's an automatically generated page, so the structure is uniform. If UPS change their page layout, you'd have to adjust the regex, but you'd also have to adjust your xml parser in that case.

Tim Pietzcker 2009-11-20 07:08:37

There should be a space between `Ref#:` and `(.{9})`. In addition, I would replace both `(.{9})` and `(.{27})` with `[^<]+`, to just read until the next tag (end) instead of a fixed number of characters.

Heinzi 2009-11-20 07:12:14

Tim / Heinzi exactly what I was looking for!Tim, your spot on for the reasons listed above. If there is a better way of doing this I would be more than willing to listen, but I believe this should have me up and runnning!

Travis Walker 2009-11-20 07:21:38

This got a downvote? Sheesh...

Heinzi 2009-11-20 07:23:38

Thanks Heinzi, I added your example. Testing in RegexBuddy on a string there the second marker doesn't match, the engine takes over 25.000 steps to fail using `[^<]+` (compared to about 2500 using `.{9}` or `[^<]++`Downvote's to be expected from the "Don't do anything on HTML with regexes" dogmatists :)

Tim Pietzcker 2009-11-20 07:28:30

s/there/where/g

Tim Pietzcker 2009-11-20 07:29:03

Heinzi, Does the edit from the original post look correct? Or is there a better way to pull out the values?

Travis Walker 2009-11-21 06:51:47

@Travis: I created a separate answer for this (because comments are not good for posting source code).

Heinzi 2009-11-21 20:19:53

Answer 3

+1 A:

Regarding your updated question about pulling out the values:

For Each match As Match In Regex.Matches(input, pattern)
    Dim ref As String = match.Groups(1).Value
    Dim track As String = match.Groups(2).Value

    ' replace with insert to db... this is for testing.
    MsgBox("Ref: " + ref + vbCrLf + "Confirmation: " + track)
Next

(untested)

Heinzi 2009-11-21 17:36:40

ansaurus

tags:

views:

answers:

Get a 2 String array from HTML .... using regex?

related questions