Question

Possible Duplicate:
How can I declare a thousand separator in read.csv?

I actually have a solution for this problem, but I am curious if there is a better way to do what I was trying to do.

I scraped some data from the majorleaguesoccer.com and read it into R using

mls.reg.tmp <- read.table("../data/mls_reg_season_20100812.csv",
                          header = F, sep = ";")

Note that I used sep = “;” because some of the attendance figures where in the thousands on the websites and I scraped “as is”, e.g.,

> str(mls.reg.dat$a_tot)
 Factor w/ 164 levels " 166,060"," 171,282",..: 132 45 159 153 46 160 
158 148 150 98 ...

In hindsight, I should’ve just removed the commas in python in the pre-processing step of this project. I should also point out that there were some text fields in the data set as well.

> str(mls.reg.dat$team)
 Factor w/ 20 levels "Chicago Fire",..: 4 9 19 11 3 10 13 16 5 6 ...

Given that I want to use the attendance data as a numeric value, I converted using as.numeric and gsub. As an example in a call to ggplot:

ggplot(data = mls.reg.dat, aes(x = as.numeric(gsub(",", "", 
  mls.reg.dat$a_tot)), y = sog)) + geom_point() + 
  facet_wrap(~ team)

Question: Is this the most efficient way of working with data such as this? Or is there a specialized function for doing something along these lines?

I’m posting the question here because I spent quite a bit of time (> 30 min) just working in this simple solution and thought that others might benefit from this as well.

Answer 1

I am not aware of any specialised function, but you could do it directly when you read the data.

  data <- read.table(...)
  data$someColumn <- as.numeric(gsub(",", "", data$someColumn))

Any subsequent call can be made using data$someColumn, without need of further conversion (and easier-to-read code)

EDIT: seems to be duplicate of “How can I declare a thousand separator in read.csv?”