CODEWITHSUNDEEP
rif-statement
Adam
Adam

Reputation: 1147

if statement error, I'm missing something simple

G'day Everyone,

I am really sorry I am coming here for this problem but I am missing something very basic here. I have downloaded the ascii file from here. I have sub-setted this data to include only the columns and rows I am interested in. Now all I want to do is replace the short 'state' names with long-form state names. I need the long-form state names to feed into another function I have written.

If you can't be bother downloading the file I'll try to roughly reproduce the data.frame.

    # If data file was downloaded.
    gazet <-   read.table('file location', sep = ',')
    gazet <- gazet[,c(2:6, 8, 10, 14)]
    colnames(gazet) <- c('recID', 'authID', 'state', 'name', 'feature',
                         'postcode', 'lon', 'lat')

    towns <- gazet[which(gazet$feature %in% c('LOCB', 'SUB', 'URBN', 'POPL')
                 & !gazet$state %in% c('JBT', 'NFK', 'MCD', 'HRD', 'AAT', 'N/A')),]

    # If you don't want to download the file.
    towns <- data.frame('recID' = c('G62', 'G63', 'G64'), 
                        'state' = c('QLD', 'SA', 'VIC'),
                        'name'  = c('Balgal Beach', 'Balhannah', 'Ballan'),
                        'feature' = c('POPL', 'POPL', 'POPL'))

    # Using whichever towns you prefer, first I change state to character
    # so that it can be changed.
    towns$state <- as.character(towns$state)

    state_longform <- function(x){
      for (i in 1:length(x$state)){
        if (x[i, 'state'] == 'VIC'){
           x[i, 'state'] <- 'Victoria'
        } else{ x[i, 'state'] == 'NA'}
      }
    } 

    state_longform(towns)
    head(towns)

My function doesn't work, yet if I do each component outside of the function it does.

    towns[1, 'state'] == 'VIC'
    towns[3, 'state'] == 'VIC'

    towns[3, 'state'] <- 'Victoria'

I plan for the full function will change all of the state names to their longform (e.g. 'QLD' to 'Queensland', etc), but I started with a shorter function to make sure it works. Which it doesn't :-(.

I am sure there is something really stupid that I am missing and I am really sorry but I haven't gotten anywhere in a long time. Is stackoverflow the right place for questions when you know you are being an idiot or is there a better place for this sort of simple question?

Thanks lots for your help.

Regards, Adam

Upvotes: 0

Views: 74

Answers (2)

Rorschach
Rorschach

Reputation: 32416

Could you just do this instead?

towns[towns$state == "VIC", ]$state <- "Victoria"

If you want to swap ALL the short names with long names do something like this, but make sure that index of corresponding long and short names are the same in the longs and shorts vectors:

longs <- c("Victoria", "Queensland", "New South Wales", "South Australia")
shorts <- c("VIC", "QLD", "NSW", "SA")
towns$state <- longs[match(towns$state, shorts)]

> head(towns)
      recID authID           state         name feature postcode      lon
33704   G62     GA      Queensland Balgal Beach    POPL     4816 146.4083
33705   G63     GA South Australia    Balhannah    POPL     5242 138.8257
33706   G64     GA        Victoria       Ballan    POPL     3342 144.2289
33707   G65     GA        Victoria     Ballarat    POPL     3353 143.8626
33708   G66     GA New South Wales      Ballina    POPL     2478 153.5654
33709   G67     GA New South Wales    Balranald    POPL     2715 143.5633

Upvotes: 1

Christopher Louden
Christopher Louden

Reputation: 7582

You need to return the changed values from the function.

state_longform <- function(x){
  for (i in 1:length(x$state)){
    if (x[i, 'state'] == 'VIC'){
       x[i, 'state'] <- 'Victoria'
    } else{ x[i, 'state'] == 'NA'}
  }
  return(x)
}

You could also replace return(x) with just x on that line and the function would return the value, but I think explicate returns make the code more readable.

Upvotes: 2

Related Questions