| Re: Dataminer? [message #41228 is a reply to message #31768] |
Fri, 15 October 2004 06:09  |
nando
Messages: 1
Registered: October 2004
|
Junior Member |
|
|
Hi Michael.
I'm using DM by a number of months.
It is surely true your code can *appear* easier, but you have to consider
your performance constraints (if any).
For example, if you need to query a large number of scalar values, e.g. they
are stored in the DB as a field of a record, the DB returns an object
(IDLdbRecordset) and you have to loop on that via a number of methods as
hereafter:
myData = dblarr(N)
For i=0, n-1 do begin ;alternatevely a while statement checking for the err
on methode move execution, if you don't know the number of elements
err = obj->move(/next);go to the next record. obj is your
record-set after a query ()
record = obj->getRecord()
myData[i] = record.data
endfor
It is clear that if you have a large number of data, this approach is very
time expensive.
On the other side, I think it is a very good add-on for interactive aspects
when no large dataset must be treated. An estimation of 'large' does not
make sense (platform, code effectiveness etc.).
If you already have your database up, it is very simple to write a little
procedure to test the performance on you system.
Using a trial license of idl you can test this in 15mins.
Have a good work,
Nando.
"Michael Wallace" <mwallace.no.spam@no.spam.swri.edu.invalid> ha scritto nel
messaggio news:10mnvbh1v8g3404@corp.supernews.com...
> Has anyone used the Dataminer add-on to IDL? If so, what did you think
> of it? We're looking into buying it, but just wondering if it's really
> worth it. In our current setup, we interact with our databases in
> external code and when IDL is needed to plot or do something, we spawn
> it into action. Some tasks *appear* that they might be a little easier
> if all the code could be kept in IDL. Any experiences? Lessons
> learned? War stories?
>
> -Mike
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
