Home » Forums » screen-scraper » screen-scraper public support

Referencing results in extractor patterns

I'm trying to scrape one page, extract the data then use the result to search a second page and search for data next to the result from the first.

e.g.

Page 1 -

Alpha
Beta
Gamma

Page 2 -

Here is some random text Beta here is the data

To do this I've got a scrape file for each page and I can successfully extract Beta from page 1 (and hold it in TOKEN) but I'm having a problem trying to match the desired data using

~#TOKEN#~ ~@NEW_TOKEN@~

The log says I have no matches when I try to use ~@TOKEN@~ (the extractor pattern works on static data).

Is it possible to reference data this way within extractor patterns (I already reference the data in parameters).

‹ Scraping foreign language site Strip HTML ›
Cantankerous on 06/30/2008 at 10:11 am

Referencing results in extractor patterns

Alan,

I am an employee here at screen-scraper and he wanted me to reply to your post. It may not be the optimal solution but what I would do is:

1) create a general extractor pattern for the first page to catch the names of each show.
2) after each pattern application scrape the name of each show on the second page under a different token name and the description as something like ~@DESC@~.
3) on the second page after each pattern application, with a script compare the first show name to the second show name with if(show1.equals(show2)). When they do match then save the ~@DESC@~ variable as the actual description variable you want to keep.

This requires a lot of page calls but I am pretty sure it would get the job done.

Bryan

 bryanm on 07/02/2008 at 11:47 am

Call script from another script

Bryan,

Thanks for this, it's working well for the the data extraction but I've just got one more question. How do I call a script file from within another script?

I've tried

session.script( "My Export to File Script" );

and

session.scrapeFile( "My Export to File Script" );

but both throw errors. Do you have an online API where I can look this sort of info up rather than bothering you?

Alex

 Cantankerous on 07/03/2008 at 5:54 am

Referencing results in extractor patterns

Scott,

Is it possible to combine scrape files? Is this a native feature of SS (I've only got the pro edition) or does it need to be done as part of a script?

I'm trying to scrape a list page

http://www.nationaltheatre.org.uk/?lid=4099

but as the details pages are not that consistent, rather than trying to get the detail from them I'm trying to get a short show summary from

http://www.nationaltheatre.org.uk/?lid=171

but I'm not sure how to call the second page and try to match the data. The URL of the details page acts as a primary key.

Alex

 Cantankerous on 07/02/2008 at 4:11 am

Referencing results in extractor patterns

Alan,

At one time it was possible but it is no longer. Referencing variables within an extractor pattern text area using ~##~ is invalid. You'll need to use a different approach.

Try combining your scrapeable files when possible and using sub-extractor patterns and the special ~@DATARECORD@~ token.

http://community.screen-scraper.com/extractor_patterns

-Scott

 swilsonmc on 07/01/2008 at 11:10 am