regmatches woes in R/Query Editor

Fixed it! The regex could be more comprehensive and it's very step by step, but this R script converts UK postcodes matching the regex pattern to an 8 character format in the query editor.

# 'dataset' holds the input data for this script
# returns string w/o leading or trailing whitespace
trim <- function (x) gsub("\\s+|\\s+", "", x)
ws <- function(x){ 
    if(x>0){strrep(" ",x)}
    else {""}
}

pattern <- "^([Gg][Ii][Rr] 0[Aa]{2})|((([A-Za-z][0-9]{1,2})|(([A-Za-z][A-Ha-hJ-Yj-y][0-9]{1,2})|(([A-Za-z][0-9][A-Za-z])|A-Za-z][A-Ha-hJ-Yj-y][0-9][A-Za-z]?)))) ?[0-9][A-Za-z]{2}"
match <- function(x) {gregexpr(pattern , as.character(x), perl=TRUE)}
part <- function(x) {c(regmatches(x, match(x)))}

pCodeList <- part(dataset$post_code)
pCodeVector <- as.character(pCodeList)

for(i in 1:length(pCodeVector))
{
    pCodeVector[i] <- trim(pCodeVector[i])
    pCodeLength <- nchar(pCodeVector[i])
    pCodeRight <- toupper(substr(pCodeVector[i], pCodeLength - 2, pCodeLength))
    pCodeLeft <- toupper(substr(pCodeVector[i],1, pCodeLength -3))
    pCodeVector[i] <- paste(pCodeLeft, ws((8 - nchar(pCodeLeft)) - nchar(pCodeRight)), pCodeRight, sep='')
}

PCDS <- within(dataset, {PCDS = (pCodeVector)})

Development discussion warmly welcomed.

Beth

x