[Qgis-user] manageR update

[Carson Farmer]

Borys Jurgiel wrote:
> Agustin Lobo wrote:
>   
>> Not sure what you mean here, losing spatial information once in R?
>>     
>
> Sorry for my English, I was uncler in fact. Yes, losing, but only as an second 
> option when you don't need any spatial capabilities. It is faster and allows 
> export any loaded layer (see limitations I said previously).
>
> Now night is coming to an end at Carson's, so let's look for the Update ;)
>   
Sorry to disappoint Borys, but I didn't quite get the time to implement 
this last night (Thesis has taken priority this week...)
Will let you know when it's ready : )
> cfarmer wrote:
>   
>> Not a cyborg, but I did give up sleep about a month ago ;)
>>     
>
> ;-)
>
>   
>> Hmm, I don't use GRASS a lot, so I didn't implement this, but now that you
>> mention it, it's probably a good idea. Unfortunately (and again, in this
>> case I'm stuck with OGR drivers here for now) GRASS layers are read only
>> AFAIK. So maybe I'll set it up so that you can read them in, and then
>> export them to one of the other org formats...?
>>     
>
> It will be great. I think saving is not a problem - even shp and geotiff 
> should be sufficient at the start, and manageR provides much more! 
>   
Cool, well I'll try to add this at the same time then...
>   
>> This is my dream as well. I don't know C though, so I don't know how I
>> would write a function to convert say a QgsVectorLayer to a sp
>> Spatial*DataFrame? Anyone wanna give it a go?
>>     
>
> I was talking rather about future - if we want to keep QGIS as much modular as 
> we can, I think the best way may be reading and writing directly from/to 
> QgsMapLayer and leaving all the format handling to provider plugins. What do 
> you think? Maybe I don't consider all threats (memory usage, specifity of 
> QgsMapLayer, risk of layer changing by user (in other thread) during long 
> analysis process...)
>
>   
>> I think this is a very good idea, and I will implement this later tonight.
>> Look for an updated manageR with an added button to load only attributes.
>>     
>
> I'm looking at kwordclock and it's still dark at Victoria... :(
> :D
>
>   
>> Great! The more the merrier. This is a cool project, and I think it can
>> really set QGIS apart as an all-in-one spatial analysis solution!
>>     
>
> I use QGIS for few years, but what is happening now is amazing ;) With all the 
> external tools and easiness/power of Python/C, we have so huge potential, 
> that I use GRASS rarely and rarely.
>
> Cheers!