This is just a general comment, but quite a few of the performance hacks
for workspaces are rather arcane, and based on understanding how the
Workbench tool generates the underlying mapping file.  Features like
this one, which rely on something extremely ephemeral like the order you
draw the connections, are really not worth messing with and, I would
argue, should either not be exposed in Workbench, or should be made more
obvious.

The reason that I say this is that if I get hit by a bus and the person
who replaces me doesn't have this knowledge, then an act as simple as
disconnecting and reconnecting a transformer will break the workspace in
a non-obvious way.  There's a lot to be said for well-documented
workspaces (VERY pretty example by the way Mark), but this kind of
metadata should not be a requirement in a tool that is otherwise
extremely intuitive and hard to break.

It would be much better if Workbench automatically sensed the cases
where it should apply this kind of optimisation and just do it behind
the scenes.  If the problem is the link to the mapping file, perhaps
mapping files should be deprecated in favour of something more flexible.

That last sentence was a joke.  Just a little high-sticking. :)

Jason

-----Original Message-----
From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of
mark2atsafe
Sent: Friday, September 15, 2006 09:22
To: [email protected]
Subject: [fme] Re: SpatialFilter Usage

Hi Mark,
I'm not trying to go immediately on the defensive, but I think there's
probably more documentation and resources than you actually realize.
There's...

FME Fundamentals
FME Getting Started Guide
FME Help (within Workbench, Viewer etc)
FME Functions, Factories and Transformers Manual FME Readers and Writers
Manual FME Tutorial FME Training Course Manual (IMO the best!) FME
Objects/SDK examples Movies showing demos of key features fmepedia.com

...plus this group and the Safe support team.

The problem is, of course, that FME is so wide-ranging we can't provide
documentation of every possible situation. If there's anything you think
we are really lacking in then let us know since we'd probably create it.

Anyway - in this case I guess you set up your GeoDatabase as the source
and route all of the point features into the SpatialFilter as the
CANDIDATE and the polygon feature as the BASE. The test to perform would
probably be CONTAINS or WITHIN, you'd have to experiment. Then send all
the PASSED features to the same output feature type.

A couple of tips - use a Sampler transformer to cut down the number of
points entering the SpatialFilter, so that you can test your setup
without running the entire dataset. In fact in 2006GB or newer you can
use the source dataset parameters > advanced > max features to read and
set that so you aren't reading all 100 million features first go.

The other tip - and this will be really important here - is to try and
make sure the polygon features enter the transformer first, and that you
have the SpatialFilter set to "Bases First". This will save you LOADS of
time and memory, because FME won't have to hold all the points in
memory.

I'd suggest you check out fmepedia - http://tinyurl.com/mb2mg and
http://tinyurl.com/o95nu are useful, but this page -
http://www.fmepedia.com/index.php/Clipper_Example - is the best example
of this concept (called Clippers First instead of Bases First).

OK - I hope this helps. Use these two tips and you should be able to
test the workspace fairly easily and then run it on the entire dataset
with maximum efficiency.

Regards,

Mark

Mark Ireland, Senior Product Specialist
Safe Software Inc. Surrey, BC, CANADA
[EMAIL PROTECTED] http://www.safe.com
Solutions for Spatial Data Translation, Distribution and Access


Join us at the FME Worldwide User Conference Sept. 21-22, 2006 Vancouver BC 
Canada. For more information, visit www.safe.com/2006uc.   
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/fme/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/fme/join
    (Yahoo! ID required)

<*> To change settings via email:
    mailto:[EMAIL PROTECTED] 
    mailto:[EMAIL PROTECTED]

<*> To unsubscribe from this group, send an email to:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply via email to