sedona_spatial_join {apache.sedona}

Perform a spatial join operation on two Sedona spatial RDDs.

Description

Given spatial_rdd and query_window_rdd, return a pair RDD containing all pairs of geometrical elements (p, q) such that p is an element of spatial_rdd, q is an element of query_window_rdd, and (p, q) satisfies the spatial relation specified by join_type.

Usage

sedona_spatial_join(
  spatial_rdd,
  query_window_rdd,
  join_type = c("contain", "intersect"),
  partitioner = c("quadtree", "kdbtree"),
  index_type = c("quadtree", "rtree")
)

Arguments

spatial_rdd

Spatial RDD containing geometries to be queried.

query_window_rdd

Spatial RDD containing the query window(s).

join_type

Type of the join query (must be either "contain" or "intersect"). If join_type is "contain", then a geometry from spatial_rdd will match a geometry from the query_window_rdd if and only if the former is fully contained in the latter. If join_type is "intersect", then a geometry from spatial_rdd will match a geometry from the query_window_rdd if and only if the former intersects the latter.

partitioner

Spatial partitioning to apply to both spatial_rdd and query_window_rdd to facilitate the join query. Can be either a grid type (currently "quadtree" and "kdbtree" are supported) or a custom spatial partitioner object. If partitioner is NULL, then assume the same spatial partitioner has been applied to both spatial_rdd and query_window_rdd already and skip the partitioning step.

index_type

Controls how spatial_rdd and query_window_rdd will be indexed (unless they are indexed already). If "NONE", then no index will be constructed and matching geometries will be identified in a doubly nested- loop iterating through all possible pairs of elements from spatial_rdd and query_window_rdd, which will be inefficient for large data sets.

Value

A spatial RDD containing the join result.

See Also

Other Sedona spatial join operator: sedona_spatial_join_count_by_key()

Examples

library(sparklyr)
library(apache.sedona)

sc <- spark_connect(master = "spark://HOST:PORT")

if (!inherits(sc, "test_connection")) {
  input_location <- "/dev/null" # replace it with the path to your input file
  rdd <- sedona_read_dsv_to_typed_rdd(
    sc,
    location = input_location,
    delimiter = ",",
    type = "point",
    first_spatial_col_index = 1L
  )
  query_rdd_input_location <- "/dev/null" # replace it with the path to your input file
  query_rdd <- sedona_read_shapefile_to_typed_rdd(
    sc,
    location = query_rdd_input_location,
    type = "polygon"
  )
  join_result_rdd <- sedona_spatial_join(
    rdd,
    query_rdd,
    join_type = "intersect",
    partitioner = "quadtree"
  )
}