Re: Programming access to Catalog

[Adam Brin <abrin at brynmawr dot edu>]

At the Tri-Colleges, we make heavy use of scripts to manipulate data from
our III System.  while I too would *love* to see a formal api or workable
xml interface, currently we have to make due.

For those using expect, I've found that the Perl module (Expect.pm) is
superior to TCL version.  It seems to get around some of the matching
issues we've had with the File and FTP interfaces.

- adam

 On Tue, 11 Oct 2005, Chad Dorton
wrote:

>
> Quoting "Hope, Ben (NIH/OD/ORS)" <hopeb at exchange dot nih dot gov>:
>
> > Aside from the Innovative customers that have the Oracle version of the
> > catalog system, we would like to know how other sites are accessing the
> > catalog programmatically or otherwise to get the data in external systems.
> > An example would be pulling records from Innovative catalog into a SQL
> > database or pulling records out to display new journals lists, or pulling
> > data out to display online resources.
> >
> > If you are writing programs, how are you actually accessing the data on the
> > Innovative system?
> >
> > Thank you.
> >
> >
>
> <editorial comment>
> I, for one, think that III is doing a giant disservice to non-Oracle sites by
> not creating SQL based or other types of programming APIs into various
> database
> functions.  It really would make their system much more powerful.
> </editorial comment>
>
> That being said, I program with the III system in 2 ways.  There may be
> others,
> but these are the two I have experience with.
>
> 1. Create lists - The search criteria in Create Lists, while not
> omnipotent, is
> quite rich and powerful, especially given that you can specify search criteria
> using regular expressions.  Once the lists of records have been created, just
> about any field you need from said records can be exported to a file.
>
> Once you have the data in a file, the sky is the limit.  You can pretty much
> munge the data with any language you like.  I tend to use Perl exclusively for
> this sort of thing, but only because I know it and (mostly) love it.  There's
> always lots of talk on this listserv about people using Excel and Access to
> work with the data, if that's your cup of tea.
>
> One drawback of this method: it isn't dynamic.  At all.  Create Lists can
> potentially take hours to run, especially on large databases.
>
> 2. Expect - I tend to write scripts that automate Create Lists
> overnight, export
> them, and send them to Perl scripts for further processing.  Expect is
> a lot of
> fun, but III's text based interface can sometimes make this challenging.  A
> little knowledge of VT100 formatting codes is helpful.  I use Expect to
>   telnet (ssh) to our III server, literally type what a patron or staff member
> would, suck up the output, and then process it. Note that this only works with
> text based (telnet/ssh) access, so if you are Millennium only, this won't work
> for you.
>
> Like I said, there may be better ways to skin this cat, but those are
> the ones I
> am familiar with.
>
> --
> Chad Dorton
> Network Specialist
> Salt Lake City Public Library
> 210 East 400 South
> Salt Lake City, UT 84111
> 801-524-8282
> cdorton at slcpl-removethis.org
> --
> This message was distributed through the Innovative Users Group INNOPAC list
> Public replies:  INNOPAC at innopacusers dot org
> Update your subscription options: http://innopacusers.org/mailman/listinfo/innopac
>