SpatVector input implemented but not yet available?

[AMBarbosa]

Hi, here you present a couple of examples using a "SpatVector" (brazil_spat) as input, but when I try that myself, I get Error in (function (classes, fdef, mtable) : unable to find an inherited method for function ‘exact_extract’ for signature ‘"SpatRaster", "SpatVector" (both with the CRAN and with the GitHub version of 'exactextractr'). Will the "SpatVector" implementation become available soon? It will be really useful! Cheers

[dbaston]

Those examples are comparing the performance of exactextractr::exact_extract to terra::extract. I don't have any plans to add support for SpatVector but would be happy to accept a contribution that does so.

[AMBarbosa]

My bad, I didn't notice some examples were with extract and others with exact_extract ;) But minimal implementation could be something so simple as coercing to 'sf' at the beginning of the function:

if (inherits(y, "SpatVector"))  y <- st_as_sf(y)

... and then adding "SpatVector" to the help file @param y

[dbaston]

I've been on the fence about that, because I'm wary of silently copying the full input dataset. But I guess it's no different than what I'm doing to "support" sp types. Do you think a warning would be appropriate, or is it just not that big a deal for most users?

[AMBarbosa]

I don't think it is warning-worth, but it might indeed be a good idea to mention in the help file @param y that 'Spatial*' and 'SpatVector' inputs will be converted to 'sf'.

[fabeit]

I am not sure if it's now possible to do exact_extract(rast,vect,fun=sum) ?

[dbaston]

I am not sure if it's now possible to do exact_extract(rast,vect,fun=sum) ?

You would need to do exact_extract(rast, st_as_sf(vect), fun = 'sum').