[Geoserver-users] What's the bounding box for?

GeoServer GeoServer User
1

Hi all,

What is the bounding box of a layer (specified in the admin console feature type page) actually used for?

What happens if a feature is added to the table outside the bounding box?

Does geoserver attempt to render areas outside the bounding box if requested from WFS or WMS?

What are the implications of the bounding box being slightly wrong / wildly wrong?

The user manual (http://docs.codehaus.org/display/GEOSDOC/4+FeatureType+Config ) seems to suggest that approximate values are ok so I’m guessing the answers to my questions are that it doesn’t matter too much but in testing of editing Oracle data in uDig via WFS having the bounding box (wildly) wrong broke it all completely.

cheers,

Tom

2

Tom,
ISO19142 (WFS 1.1.0) states:
“The element is used to indicate the edges of an enclosing rectangle in decimal degrees of latitude and longitude in WGS84. Its purpose is to facilitate geographic searches by indicating where instances of the particular feature type exist. Since multiple LatLongBoundingBoxes can be specified, a WFS can indicate where various clusters of data may exist.”

Please, note that WFS 1.1.0 contains element.
WFS 1.0.0 contains element; geoServer 1.5.1 implements WFS 1.0.0.
[ http://www.nabble.com/Re%3A-WFS-LatLongBoundingBox-p11139257.html]

pg

On 7/13/07, Tom (JDi Solutions) < tom.dean@anonymised.com> wrote:

Hi all,

What is the bounding box of a layer (specified in the admin console feature type page) actually used for?

What happens if a feature is added to the table outside the bounding box?

Does geoserver attempt to render areas outside the bounding box if requested from WFS or WMS?

What are the implications of the bounding box being slightly wrong / wildly wrong?

The user manual ( http://docs.codehaus.org/display/GEOSDOC/4+FeatureType+Config ) seems to suggest that approximate values are ok so I’m guessing the answers to my questions are that it doesn’t matter too much but in testing of editing Oracle data in uDig via WFS having the bounding box (wildly) wrong broke it all completely.

cheers,

Tom

This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/

Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users

–
Piergiorgio Cipriano
pg.cipriano@anonymised.com

3

Tom (JDi Solutions) wrote:

Hi all,

What is the bounding box of a layer (specified in the admin console feature type page) actually used for?

I believe the only thing it is used for is reporting in the Capabilities documents.

What happens if a feature is added to the table outside the bounding box?

Nothing. In the future I'd like to hook up the WFS transaction stuff to have part of its processing update the bounding box (if the admin chooses that option), but that's a bit off.

Does geoserver attempt to render areas outside the bounding box if requested from WFS or WMS?

Yes. Like I said, it's just used in the capabilities documents.

What are the implications of the bounding box being slightly wrong / wildly wrong?

From the GeoServer point of view, none, except your capabilities will report the wrong bounding box. The client may be a different story

The user manual (http://docs.codehaus.org/display/GEOSDOC/4+FeatureType+Config) seems to suggest that approximate values are ok so I'm guessing the answers to my questions are that it doesn't matter too much but in testing of editing Oracle data in uDig via WFS having the bounding box (wildly) wrong broke it all completely.

Hmmmm... That's probably more of a question for the uDig list. I mean, a WFS is supposed to report the right bounds, but I don't know why a good client would rely on that for anything more than like an overview. If you're actually downloading and editing features I'd imagine it should use the bounding boxes of the features returned by the WFS request, since it's included in all responses. But I don't know how the uDig code actually works (even though we use it too for cascading WFS, but we've never fully supported it so haven't looked at it in depth).

best regards,

Chris

cheers,

Tom
!DSPAM:4005,469792ae136981137850744!

------------------------------------------------------------------------

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/

!DSPAM:4005,469792ae136981137850744!

------------------------------------------------------------------------

_______________________________________________
Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users

!DSPAM:4005,469792ae136981137850744!

4

Tom (JDi Solutions) ha scritto:

Hi all,

What is the bounding box of a layer (specified in the admin console feature type page) actually used for?

What happens if a feature is added to the table outside the bounding box?

Does geoserver attempt to render areas outside the bounding box if requested from WFS or WMS?

Yes, it does, provide you ask for the areas outside :slight_smile:

What are the implications of the bounding box being slightly wrong / wildly wrong?

Theoretically there should not be any, besides that all standard clients would be thrown off and display the wrong starting area.

The user manual (http://docs.codehaus.org/display/GEOSDOC/4+FeatureType+Config) seems to suggest that approximate values are ok so I'm guessing the answers to my questions are that it doesn't matter too much but in testing of editing Oracle data in uDig via WFS having the bounding box (wildly) wrong broke it all completely.

This may be an Oracle index issue, but I'm not sure. I still have your
sample data, which bbox did you provide that triggered the unexpected
behaviour?

Cheers
Andrea

5

What are the implications of the bounding box being slightly wrong /
wildly wrong?

Theoretically there should not be any, besides that all standard clients
would be thrown off and display the wrong starting area.

The user manual
( http://docs.codehaus.org/display/GEOSDOC/4+FeatureType+Config
<http://docs.codehaus.org/display/GEOSDOC/4+FeatureType+Config>) seems
to suggest that approximate values are ok so I’m guessing the answers to
my questions are that it doesn’t matter too much but in testing of
editing Oracle data in uDig via WFS having the bounding box (wildly)
wrong broke it all completely.

This may be an Oracle index issue, but I’m not sure. I still have your
sample data, which bbox did you provide that triggered the unexpected
behaviour?

It broke completely when I entered the bounding box in British National Grid (instead of lat/lon) so instead of

-0.585202579443923     51.9831281911821
0.606712760960056
      52.8767813091911

I had

497152.89 : 575404.54, 232546.52 : 334177.69

. If the bounding box doesn’t really make any difference at the moment why was that a problem?
Is it because the values returned by the getCapabilities upset uDig because they’re outside the possible bounds?

Either way, it doesn’t actually matter too much now, I’ve got what I need, i.e. basically it doesn’t matter too much if they’re slightly out.

thanks,

Tom

6

On 7/13/07, Piergiorgio Cipriano <pg.cipriano@anonymised.com> wrote:

Tom,
ISO19142 (WFS 1.1.0) states:
“The element is used to indicate the edges of an enclosing rectangle in decimal degrees of latitude and longitude in WGS84. Its purpose is to facilitate geographic searches by indicating where instances of the particular feature type exist. Since multiple LatLongBoundingBoxes can be specified, a WFS can indicate where various clusters of data may exist.”

That’s interesting, Geoserver only supports one bounding box though right? but in future I could read the getCapabilities doc to check if any requested map tile was within the bounding boxe s set for a specific layer to determine whether it’s worth processing and caching it.

I’m coming at this from the angle that I want to be able to prevent the server doing unnecessary processing if a user scrolls way off the edge of our map data as currently my Open Layers + Geoserver implementation would continue rendering and caching tiles as long as the user keeps scrolling, regardless if there was actually any data to view.

Please, note that WFS 1.1.0 contains element.
WFS 1.0.0 contains element; geoServer 1.5.1 implements WFS 1.0.0.
[ http://www.nabble.com/Re%3A-WFS-LatLongBoundingBox-p11139257.html]

pg

On 7/13/07, Tom (JDi Solutions) < tom.dean@anonymised.com> wrote:

Hi all,

What is the bounding box of a layer (specified in the admin console feature type page) actually used for?

What happens if a feature is added to the table outside the bounding box?

Does geoserver attempt to render areas outside the bounding box if requested from WFS or WMS?

What are the implications of the bounding box being slightly wrong / wildly wrong?

The user manual ( http://docs.codehaus.org/display/GEOSDOC/4+FeatureType+Config ) seems to suggest that approximate values are ok so I’m guessing the answers to my questions are that it doesn’t matter too much but in testing of editing Oracle data in uDig via WFS having the bounding box (wildly) wrong broke it all completely.

cheers,

Tom

This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/

Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users

–
Piergiorgio Cipriano
pg.cipriano@anonymised.com

7

Tom (JDi Solutions) ha scritto:

On 7/13/07, *Piergiorgio Cipriano* <pg.cipriano@anonymised.com <mailto:pg.cipriano@anonymised.com>> wrote:

    Tom,
    ISO19142 (WFS 1.1.0) states:
    "The <WGS84BoundingBox> element is used to indicate the edges of an
    enclosing rectangle in decimal degrees of latitude and longitude in
    WGS84. Its purpose is to facilitate geographic searches by
    indicating where instances of the particular feature type exist.
    Since multiple LatLongBoundingBoxes can be specified, a WFS can
    indicate where various clusters of data may exist."

That's interesting, Geoserver only supports one bounding box though right? but in future I could read the getCapabilities doc to check if any requested map tile was within the bounding boxe s set for a specific layer to determine whether it's worth processing and caching it.

I'm coming at this from the angle that I want to be able to prevent the server doing unnecessary processing if a user scrolls way off the edge of our map data as currently my Open Layers + Geoserver implementation would continue rendering and caching tiles as long as the user keeps scrolling, regardless if there was actually any data to view.

If you look at the GeoServer preview functionality you'll notice you
cannot escape away from the declared bbox. This is because I locked
OpenLayers to that bbox using maxExtents (if I remember properly
the name of that property).

Cheers
Andrea

8

Tom (JDi Solutions) ha scritto:

It broke completely when I entered the bounding box in British National Grid (instead of lat/lon) so instead of

-0.585202579443923 51.9831281911821
0.606712760960056
      52.8767813091911

  I had

497152.89 : 575404.54, 232546.52 : 334177.69

Hmm... these are swapped, that is, first upper left, then
lower right? I have to try this and see what happens.
Maybe we're missing a basic bbox validity check.

Cheers
Andrea

9

ignore which way round they are, they were oriented correctly, I assumed it was just the fact that 6 digit co-ordinates are not possible in lat/lon?

On 7/14/07, Andrea Aime <aaime@anonymised.com> wrote:

Tom (JDi Solutions) ha scritto:

It broke completely when I entered the bounding box in British National
Grid (instead of lat/lon) so instead of

-0.585202579443923 51.9831281911821
0.606712760960056
52.8767813091911

I had

497152.89 : 575404.54, 232546.52 : 334177.69

Hmm… these are swapped, that is, first upper left, then
lower right? I have to try this and see what happens.
Maybe we’re missing a basic bbox validity check.

Cheers
Andrea