GS wrote:
I note that in Windoze, that "\" is used,and on the web, "/" is used.
For clarity.., a Windows file path is not the same as a URL.
Windows file paths allow spaces; URLs do not.
* Incorrect! See -------------------------vvv
https://www.nsncenter.com/NSNSearch?...r&PageNumber=5
Windows path delimiter is "\"; Web path delimiter is "/".
The function URLDownloadToFile() downloads szURL to szFilename.
* IF and when it works.
Once downloaded, szFilename needs to be opened, parsed, and result
stored locally. Then the next file needs to be downloaded, parsed, and
stored. And so on until all files have been downloaded and parsed.
* This i knew from the git-go; nice to be clarified.
Since the actual page contents comprise only a small portion of the
files being downloaded, there size should be considerably smaller after
parsing. If you extract the data (text) only (no images) and save this
to a txt file you should be able to 'append' to a single file which
would result in occupying far less disc space. (For example, pg5 is less
than 1kb) I suspect, though, that you need the image to identify the
item source (Raytheon, RCA, Lucent Tech, NAWC, MIL STD, etc) because
this info is not stored in the image file metadata. Otherwise, the txt
file after parsing pg5's text is the following 53 lines:
NSN 5960-00-509-3171
5960-00-509-3171
ELECTRON TUBE
NSN 5960-00-569-9531
5960-00-569-9531
ELECTRON TUBE
NSN 5960-00-553-3770
5960-00-553-3770
ELECTRON TUBE
NSN 5960-00-682-8624
5960-00-682-8624
ELECTRON TUBE
NSN 5960-00-808-6928
5960-00-808-6928
ELECTRON TUBE
NSN 5960-00-766-1953
5960-00-766-1953
ELECTRON TUBE
NSN 5960-00-850-6169
5960-00-850-6169
ELECTRON TUBE
NSN 5960-00-679-8153
5960-00-679-8153
ELECTRON TUBE
NSN 5960-00-134-6884
5960-00-134-6884
ELECTRON TUBE
NSN 5960-00-061-8610
5960-00-061-8610
ELECTRON TUBE
5960-00-067-9636
ELECTRON TUBE
The file size is 711 bytes, and lists 11 items. Note the last item has
no image and so no filler text (NSN line). This inconsistency makes
parsing the contents difficult since you don't know which items do not
have images.
* I think you may have pulled the info from what you saw on that page,
and not from the source.
In one of my responses, i gave QBASIC code for parsing, and as i
remember, there were about 7760 lines of junk before one sees <a
href="/NSN/5960; which gives the full NSN code.
Use of that allows one to get the second URL, eg:
https://www.nsncenter.com/NSN/5960-00-754-5782 NO image reliance at all.
There are 11 entries per page,and no inconsistencies with my method
of search in the page.
If you copy/paste pg5 into Excel you get both text and image. You could
then do something to construct the info in a database fashion...
* That would only make things more difficult. A copy to a local file is
sufficient for a simple parsing as described here and elsewhere in this
thread.
Col Headers:
Source :: PartNum :: Description
..and put the data in the respective columns. This seems very
inefficient but is probably less daunting than what you've been doing
manually thus far. Auto Complete should be helpful with this, and you
could sort the list by Source. Note that clicking the image or part# on
the worksheet takes you to the same page as does clicking it on the web
page. In the case of pg5, the data will occupy 11 rows.
* Manual: Right click, select View Page Source, Save as to HD by
changing Filetype from HTM to TXT and changing fiiename to add page
number (013 for example).
Like i said,parsing of that file is simple and easy; getting 35 pages
copied that way did not take long, but there are 999 of them...
Seems like your approach is the long way; -I'd find a better data source
myself! Perhaps subscribe to an electronics database utility (such as my
CAD software would use) that I can update by downloading a single db
file<g
* I have asked, and received zero response.