Problem: I would like to write a vectorised version of 'findClosestOutputArea' below, so that I could do something like: incidents$outputAreaGridId <- findClosestOutputAreaVectorised(incidents$davetazLongitude, incidents$davetazLatitude)$gridId 'spDistsN1' comes from package sp http://cran.r-project.org/web/packages/sp/sp.pdf and calculates the …

head(incidents)

        date     time   incidentGroup    borough         ward firstPumpTime firstPumpStation
1 2009-01-01 00:00:37 Special Service Wandsworth    Battersea           319        Battersea
3 2009-01-01 00:03:00            Fire    Enfield     Edmonton           308         Edmonton
4 2009-01-01 00:04:27            Fire Hillingdon   Hillingdon           210       Hillingdon
5 2009-01-01 00:05:39            Fire  Islington     Holloway           233         Holloway
6 2009-01-01 00:06:03     False Alarm     Camden Kentish Town           172     Kentish Town
7 2009-01-01 00:12:31 Special Service  Greenwich    Lee Green           522   East Greenwich
  latitude   longitude davetazLatitude davetazLongitude
1 51.47548 -0.14733791          51.475          -0.1485
3 51.63247 -0.06724327          51.632          -0.0675
4 51.53378 -0.44635059          51.533          -0.4470
5 51.55132 -0.10963551          51.551          -0.1110
6 51.55079 -0.13273866          51.550          -0.1335
7 51.46174  0.03063764          51.461           0.0300

head(outputAreas)

  gridId longitudeCentre latitudeCentre davetazArea
1       1122313323          -0.504       51.11170    926.4539
2       1122313324          -0.552       51.11170    926.4539
3       1122313331          -0.552       51.14050    925.8764
4       1122313332          -0.504       51.14050    925.8764
5       1122313333          -0.504       51.16935    928.5105
6       1122313334          -0.552       51.16935    812.0626

The non-vectorised function

findClosestOutputArea <- function (incident) {
    distances <- spDistsN1(data.matrix(outputAreas[, c('longitudeCentre', 'latitudeCentre')]), matrix(c(incident$davetazLongitude, incident$davetazLatitude), nrow = 1, ncol = 2, byrow = TRUE), longlat = TRUE)
    outputAreas[match(min(distances), distances), ]
}