R help - Oct 2011 - Issue with read.csv treatment of numerics enclosed in quotes (and a confession)

Dear Help-Rs,
 
I've been dealing with this problem for some time, using a work-around to
deal with it. It's time for me to come clean with my ineptitude and seek a
what has got to be a more streamlined solution from the Help-Rverse.
 
I regularly import delimited text data that contains numerics enclosed in quotes
(e.g., "00765288071").  Thing is, for some of these data, I need to
keep the values as "character" class within the data frame (that is to
say the leading zeros are important and I would like them to stay).  Here is an
example of the code I would use to read an example dataset in question:
 
mydata <- read.csv("~/mydata.csv", quote =
"\"'")
 
The problem is, when R reads the data and converts them into a data frame,
inevitably, R ignores the quotes around values like the above, and reads them in
as "numeric".  So R strips the valuable leading zeros and converts my
"00765288071" to 765288071.  I've developed a work-arounds to this
involving the use of the following:
> whatIneed <- "00000000000"
> whatIgot <- 765288071
> whatIgot <- as.character(whatIgot)
> substr(whatIneed, 1+nchar(whatIneed)-nchar(whatIgot), nchar(whatIneed))
<- whatIgot
> whatIneed
[1] "00765288071"

My question is, am I missing something in how I'm writing my read.csv
statement that would indicate to R that numerics enclosed in quotes should be
read and imported as characters and not converted to numerics???
	[[alternative HTML version deleted]]

Hi Chris,

Yes, you're missing something: the colClasses argument to read.csv.

Given a tiny little csv file that looks like this:

1,2,3,"01234"
4,5,6,"00011"
7,8,0,"00000"
> testdata <- read.csv("testdata.csv", header=FALSE,
colClasses=c(NA, NA, NA, "character"))
> testdata
  V1 V2 V3    V4
1  1  2  3 01234
2  4  5  6 00011
3  7  8  0 00000
> str(testdata)
'data.frame':	3 obs. of  4 variables:
 $ V1: int  1 4 7
 $ V2: int  2 5 8
 $ V3: int  3 6 0
 $ V4: chr  "01234" "00011" "00000"

That should do what you want.

Not that you should need it, but sprintf() is a neater way to pad out
numeric to character values:
> sprintf("%05d", 12)
[1] "00012"
> sprintf("%05d", 1223)
[1] "01223"
> sprintf("%07d", 12)
[1] "0000012"
>

Hope that solves your problem,
Sarah

On Wed, Oct 5, 2011 at 8:24 PM, Chris Conner <connerpharmd at yahoo.com>
wrote:
> Dear Help-Rs,
>
> I've been dealing with this problem for some?time, using a work-around
to deal with it. It's time for me to come clean with my ineptitude and seek
a what has got to be a more streamlined solution from the Help-Rverse.
>
> I regularly import delimited text data that contains numerics enclosed in
quotes (e.g., "00765288071").? Thing is, for some of these data, I
need to keep the values as "character" class within the data frame
(that is to say the leading zeros are important and I would like them to stay).?
Here is an example of the code I would use to read an example dataset in
question:
>
> mydata <- read.csv("~/mydata.csv", quote =
"\"'")
>
> The problem is, when R reads the data and converts them into a data frame,
inevitably, R ignores the quotes around values like the above, and reads them in
as "numeric".? So R strips the valuable leading zeros and converts my
"00765288071" to 765288071.? I've developed a work-arounds to this
involving the use of the following:
>
>> whatIneed <- "00000000000"
>> whatIgot <- 765288071
>> whatIgot <- as.character(whatIgot)
>> substr(whatIneed, 1+nchar(whatIneed)-nchar(whatIgot), nchar(whatIneed))
<- whatIgot
>> whatIneed
> [1] "00765288071"
>
> My question is, am I missing something in how I'm writing my read.csv
statement that would indicate to R that numerics enclosed in quotes should be
read and imported as characters and not converted to numerics???
>

-- 
Sarah Goslee
http://www.functionaldiversity.org