[R] re ading transposed data in read.csv

[Vitalie S.]

On Sat, 15 Aug 2009 09:20:04 +0200, Dieter Menne  
<dieter.menne at menne-biomed.de> wrote:

>
>
>
> sandsky wrote:
>>
>> I have a data set (csv); e.g.,
>>
>> ID  samp1  samp2 samp3 samp4
>> G1       23        32      12      87
>> G2       85        45      49      76
>> G3       12        46      39      28
>> G4       73        26      18      13
>>
>> and read it:
>>
>> data1<-(read.csv("Datafolder/rawdata.csv",header=T))
>>
>> It is fine with reading as above, but, if it is read as transposed  
>> values
>> like:
>>
>> data1<-t(read.csv("Datafolder/rawdata.csv",header=T))
>>
>> ID         "G1"   "G2"   "G3"   "G4"
>> samp1   "23"    "85"    "12"   "73"
>> ...
>>
>> all values are shown in "   ".
>>
>> How to get this transposed data as numerical values?
>>
>>
>
> Because ID is not numeric, and in a data frame each column must have the
> same data type. The least common denominator is "character".
>
> Written explicitly, you have to transpose the numeric part, and paste in  
> the
> names later. This could be shortened, but the example below gives the  
> idea.
>
> Dieter
>
> d = data.frame(ID=c("G1","G2"), samp1=1:2,samp2=5:6)
> d
> t(d)
> dt = as.data.frame(t(d[-1]))
> names(dt)= d[,1]
>
>
>


As Dieter has already said elements in each of column of a data.frame must  
be of the same type. But this is probably what you really after:

> tdf <- read.table(textConnection("ID  samp1  samp2 samp3 samp4
+ G1       23        32      12      87
+ G2       85        45      49      76
+ G3       12        46      39      28
+ G4       73        26      18      13
+  "), header=T)
> library(reshape)
> recast(tdf, ...~ID)
Using ID as id variables
   variable G1 G2 G3 G4
1    samp1 23 85 12 73
2    samp2 32 45 46 26
3    samp3 12 49 39 18
4    samp4 87 76 28 13
>

see reshape and plyr packages for this kind of data manipulation.

-Vitalie.


--