fnALSourceParse ( source ; attribute )

Returns a list of values for the supplied attribute from the source code of a webpage. See - fnHTMLAttributeList

Average rating: 4.0 (37 votes) Log in to vote

Jeremy Upton
DME Supply Group
http://www.dmesupplygroup.com

Sample input:

fnALSourceParse ( "\source code of http://www.briandunning com" ; 'href" )

Sample output:

http://www.briandunning.com/
http://www.briandunning.com/consulting/ventures.php
http://www.briandunning.com/tomarket/
http://www.briandunning.com/filemaker-pro/
http://www.briandunning.com/filemaker-custom-functions/
http://www.briandunning.com/filemaker-web-viewer/
http://www.briandunning.com/consulting/
http://www.briandunning.com/training.shtml
http://www.briandunning.com/books.shtml
http://www.briandunning.com/magazines/
http://www.briandunning.com/error-codes/?source=FileMaker
http://www.briandunning.com/free-web-tools/
http://www.briandunning.com/free-filemaker-tools/
http://www.briandunning.com/personal.shtml
http://www.briandunning.com/videos/
http://www.briandunning.com/adventures/
http://www.briandunning.com/links/
etc... (not showing complete results)

Function definition: (Copy & paste into FileMaker's Edit Custom Function window)

Let ( [

//Count of occurrences of the attribute within the source code
LinkCount= PatternCount ( source; attribute );

//Position of the first occurrence of the attribute within the source code
CurrentPosition= Position ( source ; attribute ; 1 ; 1 );

//Resulting source code after everything up to the first attribute occurrence has been removed
AdjSource= Right ( source; Length ( source ) - CurrentPosition );

//Determines if the attribute is part of commented text
Comment= PatternCount ( Left ( source; CurrentPosition ); "<!--" ) > PatternCount ( Left ( source; CurrentPosition ); "-->" );

//If the attribute is not within commented text everything within the following " " is parsed
CurrentLink= If (  not Comment; 
                           Substitute ( 
                                Middle ( source ; 
                                Position ( source ; "\"" ; CurrentPosition ; 1 ) + 1 ; 
                                Position ( source ; "\"" ; CurrentPosition ; 2 ) - (Position ( source ; "\"" ; CurrentPosition ; 1 ) + 1) )
                           ; "\"" ; "" ); 
                      " ");

//Determines if there are any remaining attribute left in the AdjSource
Continue= LinkCount ≠ 0

];

//the substitute removes any blank values or double returns

Substitute ( 
     If ( Continue; 
          CurrentLink & ¶ & fnALSourceParse ( AdjSource; attribute ) ) ;
[¶ & " " & ¶ ; ¶]; [" " & ¶; ""]; [¶ & "" & ¶; ¶] )

 )

NOTICE: This function is intended to be used in conjunction with the custom function - fnHTMLAttributeList available from www.briandunning.com. It makes things much easier, however it does not require it.

This function does a simple thing. It takes the source code provided and locates every occurrence of the text (attribute) that you request. It then returns the next value within quotes (" ") following each of those occurrences in a list format. You can literally type anything for the attribute parameter, but the function really becomes useful when you enter html attributes. Here are some examples:

Entering "src" can be used to get a list of images on page.
Entering "href" will produce a fairly complete list of links on the page. As seen in the sample output.

Information like this can come in very handy for tasks like web scraping.

In my testing I found that the usefulness of the results depended greatly on the how the site was organized. For instance, a page that utilizes a lot of javascript will not produce much that you can use. However, you can fine tune the string that you want the function to look for and often times get better results. For example, entering "
If anyone out there with more knowledge of html than myself knows how this can be improved I welcome your input.

Comments

		Mark Keedy, Kenosha, Wisc. Sep 12, 2013
How would you modify this if the attribute string itself has quotation marks? I don't want all the href links from my site, just the ones that are preceded by '<a class="cssa" href='

Note: these functions are not guaranteed or supported by BrianDunning.com. Please contact the individual developer with any questions or problems.

fnALSourceParse ( source ; attribute )

Comments

Support this website.

Support this website.