C# HTML parse, regular expression or ?

I checked and cannot publish directly on the forum. I do not have XML but HTML so need HTML to data parser

http://msdn.microsoft.com/en-us/library/ms756177(VS.85).aspx

http://blogs.msdn.com/tims/archive/2007/06/13/programming-html-with-c.aspx

I hope this little information will help.

HTML document is not XML (so one article is not related to HTML), another articles works upon currenctly loaded document.
the HTML I have is on the disk drive and I need to open and parse it, that's all (no Silverlight to use on active document as another link is about)
I can have console or win application that takes input file name  and database connection string as an argument and loads that from HTML into database.
I need to match regular expresion e.g. "Area" and then take the name that apears after this text and next text which "SubArea". I create thus one record in AREA table. The I continue to road "SubArea" which might be several for area and plase corresponding number of records into database table "SubArea" etc.

HTML is a subset of XML. All rules apply.

In other words, you have a database in html ? If so, you'd have to declare table sizes and fields for each of your table-createable-tags, or create an SQL query generator that would conjure up a query depending on the fields in your table(in your file). I'm afraid I'm not aware of such a tool right now.

It's a tricky task actually. I'm sorry I'm no help. Cheers

Regular expressions are the way to go. You have to know how the data will always look and since it seems homegrown, should be fairly easily. But as piotroxp stated, hard to get you further without an example. The best way forward is to post an example file, with your non-working regexp that we can help you get working.

I attached the file.

I don't see any attachments.

Sorry, did not check, seems that html cannot be uploaded so I renamed into txt

Wow. First thing, their will be a language barrier for me, but should be doable.

2nd thing is that the HTML is non-compliant throughout. All the HTML code should be lowercase and if you are not, at least have matching tags in the same case, lots of <table width=590>.......</TABLE> type things in there. Make sure you don't do case sensitive on the regex engine.

Got a family night, but will look at this later and tomorrow.

After a few more minutes looking, I don't think one single all powerful regex will do this, not easily anyway. There are multiple formats to the data tables, which doesn't make much sense (ie, can't read it) to me.

I thought this would get me far:

<table[^>]+>(\s+<tr><td[^>]*>([^<]+)</td>(\s*<td[^>]+>([^<]+)</td>)+</tr>)+\s*</table>

but it only matches 6 of the roughly 100 tables.

Forgot EE would hack it up

<table[^>]+>(\s+<tr><td[^>]*>([^<]+)</td>(\s*<td[^>]+>([^<]+)</td>)+</tr>)+\s*</table>

                                              

1: