How hard is it to pull data from a table on a web site?

edcentricedcentric near Milwaukee, Wisconsin Icrontian
edited November 2004 in Internet & Media
I check a web site each morning.
http://www.nrc.gov/reading-rm/doc-collections/event-status/reactor-status/ps.html
Right now I need to keep a manual list of which reactors are down.
I would like to either have the data in a table for the last week so that I could see when one goes off-line. Or better yet, to create a list (exception report) of which ones have decreased in out put since the day before.
How do I write a prog to use the data?
Or am I stuck?

A good share of our business is with the nuc plants. Knowing who is having reliability problems tells me who to be calling.

Comments

  • PressXPressX Working! New
    edited November 2004
    A simple way would be to parse the data in a spread sheet and use that to highlight the on/off line plants as well as changes in output. Someone with a little VB could generate a prog to help do that for you. Especially if the page layout remains the same. Alternatively, someone with a good grip on PHP may be able to take the page and create a new one with the info you need - my php is not great so I am guessing this is possible. I am 99% sure it is but ask a professional PHP wiz.
  • GnomeWizarddGnomeWizardd Member 4 Life Akron, PA Icrontian
    edited November 2004
    see the one that says saint lucie! w00t I can take pics of it if u want!
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited November 2004
    Simple two cell-per-row table. Strip HTML around table, comma+space delimit cells, use carriage returns per each record, then import in Open Office, the spreadsheet or in write. Or, strip down to just tables with data cells, and import into OOo. Now, the first import gives you the base structure, after that simply grab the numbers and stick into adjacent columns with one column per day. Read across rows to see trends. Technically, this can be cut-n-paste after first build of table base.

    If you DO need region limiting info, then four tables, one per region, would be easiest first breakout of data. Then import data fro each region into each region's corresponding table.

    BTW, OOo files are ARCHIVES, gzip type, and the contents is XML, the other file is font and format specs. Winzip and gzip are close enough that you can extract the contents file from the OOo ARCHIVE and have your data, and you can in fact strip the surrounding HTML calls for gifs, etc., in OOo Write or Mozilla's Composer for that matter.

    PHP not needed for simple first table formatting and create with a two-cell per row table.
  • edcentricedcentric near Milwaukee, Wisconsin Icrontian
    edited November 2004
    OK, I get the jist of what is being said.
    The only problem is that I haven't programed since....'76.

    Is there a volunter to help me actually do this? About all that I can offer in return is a package of good Brats or a Cheese Head.

    I don't need the regional info. One large table would be fine. I don't need to automaticaly highlight changes. Just building a table would do.
    How automated could it be?

    Straight Man, you up for this?

    Gnome, don't take pictures. I don't want to hear that you need bail.
  • MERRICKMERRICK In the studio or on a stage
    edited November 2004
    edcentric wrote:
    OK, I get the jist of what is being said.
    The only problem is that I haven't programed since....'76.

    Sorry to make you come here for a change of topic but I am curious to know what you were programming in 1976 since I take a very layman's interest in the history of computing.

    BTW I don't like it when I need info and people start chiming in on my thread with off topics when I need (somtimes desperately) info. So if you blow me off I won't be insulted. :cool:
  • edcentricedcentric near Milwaukee, Wisconsin Icrontian
    edited November 2004
    Let's see. I was a soph at Purdue. The general computing stuff was all in Fortran (I think that we were running the UMinn version) with punch cards and run in batch. Of course if you could steal someones account info you could get higher priority. As I recal the Ag School had top access to the cental system (after CS of course). EE ran their own system. I did a little Unix and even some machine language stuff (Snowball?). The latter was all on little PDP Micro computers from DEC (11-07 and such).

    I regret that I didn't keep a box of Hollerith cards. There were some terminals where you could submit a program in batch and then step through it on the screen. We had a space war game (Tiger?) that was three boxes full of cards. We only ran it on stolen accounts.
  • MERRICKMERRICK In the studio or on a stage
    edited November 2004
    Honestly I don't recognize anything you wrote except for Perdue and Unix. I'm really only aware to post Win 3.1 history. (But I'm old enough to remember seeing those large console machines with tape reels on TV). It's cool to hear about the way things were back before computers were "Pop" culture. Thanks for the picture!
  • shwaipshwaip bluffin' with my muffin Icrontian
    edited November 2004
    How long do you need to store history data for, and how often would it need to update?
  • edcentricedcentric near Milwaukee, Wisconsin Icrontian
    edited November 2004
    This is updated daily. I need every days numbers.
    It could be broken into files for time periods. Though for ease of analysis I would like to have at least 4 months worth of data together.

    Do you have some ideas? I am interested.
  • shwaipshwaip bluffin' with my muffin Icrontian
    edited November 2004
    I have some experience doing similar things with JAVA, as I was using it to make the SMx stats page until they (they = the network/sysadmins) changed things that prevented how I was publishing the stats from working.

    It's fairly easy to get the data from the site and put it in a table or database, but the only way to make it update at a regular interval is to make it a windows scheduled task, or have it run at startup...but that can be figured out once the program is finished. If you have microsoft access, I can make the program put the data in an access database for easy viewing. Another possibility is to create a GUI for the program that lets you view individual plants history, or view by regions, etc.

    Ideally the program would be able to handle the addition of a new power plant to the page, and assorted other things that I'm sure I haven't thought of yet.

    Since this is my quarter break, and I'll be rediculously bored for a week, I've started messing with it. I've finished the grabbing and parsing of the data from the site.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited November 2004
    shwaip wrote:

    Since this is my quarter break, and I'll be rediculously bored for a week, I've started messing with it. I've finished the grabbing and parsing of the data from the site.

    Go for it, schwaip. No, Edcentric, I will be too busy for months to do this for you (we cannot get anyone to even bid on a screen cage for Lanai for us, looks like I and my brother Chris will be building the screen cage we lost during Charley, plus hanging gutter and fascia and new soffit and reworking old salvageable soffit, and maybe PAINTING part of outside of house and part of inside-- plus building my business here), but simply decided to outline roughly the process needed and let someone maybe volunteer.

    Are you using VBA plus Access, Schwaip???? An Office VBA Project could do this decently..... Including trending if it were an Access+Excel VBA app or a Crystal Reports report from Access plus Excel processing (actually, Crystal Reports Dev version could do the pull-and-calc from Access or probably a Paradox scheduled Export report file, but I only have Pro and I use Paradox and not Access, and Quattro Pro or OOo's spreadhseet and not MS Office at all right now-- in fact, last full version of MS Office I used for anything major was Office '97.).
  • shwaipshwaip bluffin' with my muffin Icrontian
    edited November 2004
    Here's a fascinating screenshot showing one day's data. If you have any comments/suggestions/requests, let me know.
    ed.JPG 148.1K
  • shwaipshwaip bluffin' with my muffin Icrontian
    edited November 2004
    now with cool-looking (TM) graph. Now I need to go through and add a couple of error messages to help in case something b0rks.

    edit: I found that the data for recent weeks was in the same format, and added that to the database. This is what a graph looks like now.

    Things to do:
    highlight those plants which have gone down in production and/or create a report that lists them.
  • shwaipshwaip bluffin' with my muffin Icrontian
    edited November 2004
    does the site not update on the weekends?
  • edcentricedcentric near Milwaukee, Wisconsin Icrontian
    edited November 2004
    shwaip, that looks like a ticket to me. Damn you are good.
    drop me an email.

    It sort of updates on weekends, sometimes the numbers are reported and sometimes not. I have never figured that out. Usually not though.
  • primesuspectprimesuspect Beepin n' Boopin Detroit, MI Icrontian
    edited November 2004
    Have I ever told anybody that I love this site? :D
  • shwaipshwaip bluffin' with my muffin Icrontian
    edited November 2004
    Program finished, and e-mail sent :D
    Have I ever told anybody that I love this site?
    Only as often as you've told someone to use memtest.
Sign In or Register to comment.