Utility to create a table used to create a quadtree spatial index of points etc.

Used by indexpoints() that in turn is used by proxistat(). It prepares a set of coordinates ready for indexing.

Usage

create_quaddata(pts, idcolname = NULL, xyzcolnames = c("x2", "z2", "y2"))

Arguments

pts

a data.frame or data.table with columns name lat and lon, one row per location (point), and any other columns are ignored.

idcolname

if NULL (default), a pointid column is created as a unique id 1:NROW(). If creating the index of blocks, idcolname is "blockid" If set to "id" it just uses that even if not unique id. indexpoints() does not directly refer to this column but index probably incorporates it.

xyzcolnames

For creating quaddata and then localtree index of blocks, this must be set to c("BLOCK_X", "BLOCK_Z", "BLOCK_Y") ??

Value

returns a data.table one row per point, columns with names that are c(xyzcolnames, idcolname)

Details

Very similar to what is used to prepare quaddata, which is used by getblocksnearby().

Note that BOTH this table and the index of it are needed in getblocksnearby() or getpointsnearby() !

For 8 million block points, this takes a couple of seconds to do, so it may be useful to store the index during a session rather than building it each time it is used. But it cannot be saved on disk because of what it is and how it works.