Ang. Re: [Robelle-l] suprtool-problem, string-related

[per.2.ostberg@kus.se]

"Barry Durand" <durand@allegro.com> skrev 2004-04-29 17:24:43:

>
>
> I know that you said the file was tab delimited but just to make
> sure are the columns fixed?   In other words, would the ID-NUMBER
> always be in (for example) columns 10 through 19?
>

hmmm, the intent (when the format was dreamed up) was to make the "columns"
variable (and have tabs as delimeters), but looking at it now, I think we
made the file both fixed-columns *and* tab-delimeted. I.e I think I could
get away with treating the columns fixed.

/per

--------------------------------------------------
Per Ostberg  / per.2.ostberg@kus.se
tel 08-58581213 / fax 08-58587710 / mob 070-4840253
LabMedicin-IT
Karolinska Universityhospital, Huddinge
C1 56 / 141 86  HUDDINGE / SWEDEN
--------------------------------------------------


> Barry Durand
>
> On 29 Apr 2004 per.2.ostberg@kus.se <per.2.ostberg@kus.se> wrote
>
> Date sent:         Thu, 29 Apr 2004 17:09:57 +0200
> From:              per.2.ostberg@kus.se
> To:                robelle-l@robelle.com
> Subject:           [Robelle-l] suprtool-problem, string-related
>
> >
> >
> >
> >
> >
> > greetings!
> > I have a tab-delimeted text-file with records looking (something) like
> > this:
> > data<tab>ID-number<tab>moreData
> >
> > the ID-number (in ascii) is formatted as xxx-xxxx-x ("dash-separated",
> > "123-4567-8" as an example), but it is also an item (numeric, an I2
item,
> > 12345678 in this example) in my TI dataset.
> > I need to split this file in two, depending on whether the ID-number
exists
> > in the dataset or not. I am thinking of building a suprtool table and
use
> > the $lookup-function, but am not able to figure out how to make
suprtool
> > understand my "dash-separated" ID-number.
> > Is this do-able in suprtool or do I have to resort to
MPE-CI-programming
> > (in combination with suprtool, of course)?
> > There's not that much data involved: the file will be a few hundred
records
> > and the data-set will probably have 10000 records.
> >
> > I'm grateful for any ideas.
> >
> > tia
> > /per
> >