set.spChkOption: Control checking of spatial object IDs

In spdep: Spatial Dependence: Weighting Schemes, Statistics

Control checking of spatial object IDs

Description

Provides support for checking the mutual integrity of spatial neighbour weights and spatial data; similar mechanisms are used for passing global verbose and zero.policy options, and for causing functions creating neighbour objects to warn if there are multiple subgraphs.

Usage

set.spChkOption(check)
get.spChkOption()
chkIDs(x, listw)
spNamedVec(var, data)
set.VerboseOption(check)
get.VerboseOption()
set.ZeroPolicyOption(check)
get.ZeroPolicyOption()
set.SubgraphOption(check)
get.SubgraphOption()
set.SubgraphCeiling(value)
get.SubgraphCeiling()
set.NoNeighbourOption(check)
get.NoNeighbourOption()
set.listw_is_CsparseMatrix_Option(check)
get.listw_is_CsparseMatrix_Option()

Arguments

check	a logical value, TRUE or FALSE
value	an integer value, initialised as 100000L, the sum of the numbers of nodes and edges in the neighbour graph
x	a vector the same length, or a two-dimensional array, or data frame with the same number of rows as the neighbours list in listw
listw	a listw object or nb object inheriting from "nb"
var	a character string or integer value for the column to be selected
data	a two-dimensional array or data frame containing var

Details

Analysis functions will have an spChk argument by default set to NULL, and will call get.spChkOption() to get the global spatial option for whether to check or not — this is initialised to FALSE, and consequently should not break anything. It can be changed to TRUE using set.spChkOption(TRUE), or the spChk argument can be assigned in analysis functions. spNamedVec() is provided to ensure that rownames are passed on to single columns taken from two-dimensional arrays and data frames.

Value

set.spChkOption() returns the old logical value, get.spChkOption() returns the current logical value, and chkIDs() returns a logical value for the test lack of difference. spNamedVec() returns the selected column with the names set to the row names of the object from which it has been extracted.

Note

The motivation for this mechanism is provided by the observation that spatial objects on a map and their attribute data values need to be linked uniquely, to avoid spurious results. The reordering between the legacy Columbus data set used the earlier publications and that available for download from the Spacestat website is just one example of a common problem.

Author(s)

Roger Bivand [e-mail]

Examples

data(oldcol)
rownames(COL.OLD)
data(columbus, package="spData")
rownames(columbus)
get.spChkOption()
oldChk <- set.spChkOption(TRUE)
get.spChkOption()
chkIDs(COL.OLD, nb2listw(COL.nb))
chkIDs(columbus, nb2listw(col.gal.nb))
chkIDs(columbus, nb2listw(COL.nb))
tmp <- try(moran.test(spNamedVec("CRIME", COL.OLD), nb2listw(COL.nb)))
print(tmp)
tmp <- try(moran.test(spNamedVec("CRIME", columbus), nb2listw(col.gal.nb)))
print(tmp)
tmp <- try(moran.test(spNamedVec("CRIME", columbus), nb2listw(COL.nb)))
print(tmp)
set.spChkOption(FALSE)
get.spChkOption()
moran.test(spNamedVec("CRIME", columbus), nb2listw(COL.nb))
tmp <- try(moran.test(spNamedVec("CRIME", columbus), nb2listw(COL.nb),
 spChk=TRUE), silent=TRUE)
print(tmp)
set.spChkOption(oldChk)
get.spChkOption()

spdep documentation built on Sept. 13, 2024, 5:07 p.m.