[R] number of patients in a hospital on a given date
Thomas Gerds
gerds at fdm.uni-freiburg.de
Wed May 14 16:18:36 CEST 2003
how about this:
x <- data.frame(patid=c("pat1", "pat2"), adm.date = c("15.03.2002","16.03.2002"),dis.date=c("18.03.2002", "17.03.2002"))
x[,2:3] <- apply(x[,2:3], MARGIN=2, FUN=strptime, format="%d.%m.%Y")
alldays <- c("2002-03-14","2002-03-15","2002-03-16","2002-03-17","2002-03-18","2002-03-19")
tmp <- table(unlist(apply(x,
1,
function(y){
beg <- match(y[2],alldays)
end <- match(y[3],alldays)
alldays[beg:end]})))
nulldays <- alldays[match(alldays,names(tmp),nomatch=0)==0]
out <- rbind(data.frame(days=c(names(tmp),nulldays),freq=c(as.numeric(tmp),rep(0,length(nulldays)))))
out[order(out$days),]
days freq
5 2002-03-14 0
1 2002-03-15 1
2 2002-03-16 2
3 2002-03-17 2
4 2002-03-18 1
6 2002-03-19 0
tomy
RINNER Heinrich <H.RINNER at tirol.gv.at> writes:
> Dear R-users!
>
> I am using R 1.7.0, under Windows XP.
>
> Having some hospital discharge data (admission date and discharge date for
> each patient), I want to get the number of patients in the hospital on a
> given date.
>
> My data look like (simple example):
>> x <- data.frame(patid=c("pat1", "pat2"), adm.date = c("15.03.2002",
> "16.03.2002"),
> dis.date=c("18.03.2002", "17.03.2002"))
>
> I can easily do a date-time conversion from the character objects:
>> x[,2:3] <- apply(x[,2:3], MARGIN=2, FUN=strptime, format="%d.%m.%Y")
>> x
> patid adm.date dis.date
> 1 pat1 2002-03-15 2002-03-18
> 2 pat2 2002-03-16 2002-03-17
>
> What I want in the end is something like a data.frame A like this:
> A
> date no.of.patients
> 2002-03-14 0
> 2002-03-15 1
> 2002-03-16 2
> 2002-03-17 2
> 2002-03-18 1
> 2002-03-19 0
>
> Or, alternatively, a data.frame B like this:
> B
> patid date.in.hospital
> pat1 2002-03-15
> pat1 2002-03-16
> pat1 2002-03-17
> pat1 2002-03-18
> pat2 2002-03-16
> pat2 2002-03-17
>
>>From this I could easily get A by using "table".
> So the trick would be to get a data.frame with one line for each day of each
> patient in the hospital - but how?
>
> I'd be happy about any ideas,
> Heinrich Rinner.
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
--
no signature
More information about the R-help
mailing list