[R] Dataframes and text identifier columns

[David Winsemius]

On Jun 29, 2014, at 5:46 AM, Brian Willis wrote:

> I am trying to incrementally add rows to an empty data frame. The  
> data has 7
> columns, the last 6 are numeric.
>
> For the first columns I would like to include a text identifier,  
> called
> ‘Case’ like Andrew, Burt, Charlie etc. that is also output to a data  
> frame –
> this is where I am having the problem
>
> The identifiers are being input from file which automatically  
> assigns them
> to a data frame
> Code:
>
> Inp_dat<- read.csv("/File/Input data.csv")
>
> out_put<-data.frame(Case=character(), StdL=numeric(), StdPP=numeric(),
> StdSE=numeric(), L=numeric(), MRPP=numeric(), MRSE=numeric(),
> stringsAsFactors=FALSE)
>
> for(i in 1:4)
> {
> if (i==1) b<-Andrew   # these assignments do not appear to ever get  
> used.
> if (i==2) b<-Burt
> if (i==3) b<-Charlie
> if (i==4) b<-Dave
>
> pr <- Inp_dat$p[i]
> SE_pr <- Inp_dat$SE[i]
> r<- Inp_dat$r[i]
> n<- Inp_dat$n[i]
> Case<- levels(Inp_dat$Case)[i])

# Since the question involves failure of this step, there woudn't  
appear to be much we could say unless you provided teh output of  
levels(Inp_dat$Case).
>
> …
>
> out_put[i,]<-data.frame(Case, stdL, stdPP, stdSE, L, PP, PP_SE)

The values stdL through PP_SE seem to come from someplace else.
>
> }
> out_put
>
>  Case     StdL          StdPP           StdSE           L
> MRPP            MRSE
> 1    1  19.466823   0.16432300   0.03137456   26.002294   0.2080145
> 0.03804692
> 2    1   2.334130    0.22566939   0.08962662    5.095703    0.3888451
> 0.08399101
> 3    1   2.588678    0.05502765   0.00454159   42.058326   0.4861511
> 0.02128030
> 4    1   7.857898    0.18457822   0.04372297    4.705487    0.1193687
> 0.01921609
>
>
> Unfortunately the Case column loses the labels Andrew, Burt …

If you were thinking that the values Andrew and Burt would come from  
that loop then it's unclear whay tha twould be so since the value of  
"b" is never used after it is defined.
>
> I’ve tried always to keep these, but are clearly doing something  
> wrong. Can
> anyone help?



David Winsemius, MD
Alameda, CA, USA