Keep the rows of a data frame that relate to another spatially

Filters x down to the rows whose geometry relates to any row of y. The columns of y are not attached, which is the difference from ga_join().

Usage

ga_filter(x, y, predicate = ga_sparse_intersects, ...)

Arguments

x, y

data frames, each with one GeoArrow geometry column

predicate

a sparse predicate, by default ga_sparse_intersects()

Value

an Arrow table holding the rows of x that matched

Details

A row of x is kept when the predicate finds it at least one match in y, so ga_filter(sites, counties, ga_sparse_within) keeps the sites that fall in any county. This is ST_Filter(), and the same answer as ga_join(x, y, predicate, left = FALSE) with the columns of y dropped and the rows of x not repeated.

See also

Other topology: ga_contains(), ga_coordinate_position(), ga_dimension(), ga_is_empty(), ga_join(), ga_relate(), ga_sparse_intersects(), ga_sparse_pairs()

Examples

nc <- as.data.frame(read_shapefile(
  system.file("shape/nc.shp", package = "sf")
))
sites <- data.frame(
  site = c("a", "b", "c"),
  geometry = geoarrow::as_geoarrow_vctr(
    ga_xy(c(-78.6, -80.8, 0), c(35.8, 35.2, 0))
  )
)

# the third site is in the Atlantic, so it goes
ga_filter(sites, nc, ga_sparse_within)
#> Table
#> 2 rows x 2 columns
#> $site <string>
#> $geometry: geoarrow.point <crs <unspecified>>
#> 
#> See $metadata for additional Schema metadata