View Single Post
  #76   Report Post  
Posted to microsoft.public.excel.programming,microsoft.public.excel
Robert Baer Robert Baer is offline
external usenet poster
 
Posts: 93
Default Read (and parse) file on the web

GS wrote:
GS wrote:
Excel macros are SO... undocumented.
Need a WORKING example for reading the HTML source a URL (say
http://www.oil4lessllc.org/gTX.htm)

Thanks.

Look here...

https://app.box.com/s/23yqum8auvzx17h04u4f

..for *ParseWebPages.zip*, which contains:

ParseWebPages.xls
NSN_5960.txt
(Blank data file with fieldnames only in line1
NSN_5960_Test.txt
(Results for 1st 20 pages)

Holy S*!
I did about 30 pages by hand; quit as rather tiresome and total pages
unknown (MORE than 999).
Never saw the fail you saw.

Difference is that you used the word "and"; technically (i think) that
should not affect results.
Also, you got items I am interested in, and after processing 503
pages, i did NOT get those.

In both cases, there were a lot of duplicate records (government data,
what else can you expect?).

In your sample, there were 73 useful records containing 43 unique
records. There may be some that i am not interested in, but there
definitely ARE those i did not find that i am interested in.


You could 'dump' the file into a worksheet and filter out the dupes
easily enough.

* Yes, i did that; getting those 43 unique records.


In my sample, there were 3782 unique records,and (better sit down),
only 15 were interesting. Crappy odds.

Hopefully, when i call them, someone that has some experience and
knowledge of how their sort criteria works, will answer the phone.
Last time i called,i got a new guy; no help other than "use Google".


Those are dynamic web pages and so are database driven. Surely there's a
repository database for this info somewhere other than NSN?

* Prolly not dynamic as NSNs do not change except possible additions on
rare occasion.
Certainly, a SPECIFIC search (so far) always gives the same results.
Look at a previous response concerning 100% manual search results,
first page only.
5960 AND REGULATOR AND "ELECTRON TUBE" About 104 results
5960 REGULATOR AND "ELECTRON TUBE" About 922 results
Totally different and neither match your results.
And your results look superior.


You have done a masterful job!
Label it !DONE! please.

Thanks a lot.


Happy to be of help; -I found the project rather interesting!