CODEWITHSUNDEEP
rif-statementsubset
mgc77
mgc77

Reputation: 113

if subset yields zero observations skip to next value in r loop

I'm plotting data and I have a loop that first finds all data corresponding to a particular ID number. Sometimes there is no data for that particular ID so I need to add a if else if statement within the loop because other wise I get an error that there is no x values for the plot

Actual Code

 df<-subset(MonthFiltered,t_all<100)
 IDunique<-unique(MonthFiltered$ID)

for (f in IDunique) {
temp<-subset(df,ID==f)
name<-paste(paste(f, "cdf", sep="-"),"png", sep=".")
png(name)
plot(ecdf(temp$t_all))
dev.off()
}

Need something like

for (f in IDunique) {
temp<-subset(df,ID==f)
#if temp obs.=0 then skip to next f
#else if 
name<-paste(paste(f, "cdf", sep="-"),"png", sep=".")
png(name)
plot(ecdf(temp$t_all))
dev.off()
}

Upvotes: 1

Views: 2188

Answers (2)

Andy Clifton
Andy Clifton

Reputation: 5066

Use NROW(temp$t_all[!is.na(temp$t_all)]) to get the length of temp$t_all that contains valid data, as follows:

df <- subset(MonthFiltered,t_all<100) 
# now loop through the unique values of ID
IDunique <- unique(MonthFiltered$ID)
for (f in IDunique) {
  temp<-subset(df,ID==f)
  # check to see if the subset actually contains non-na data
  if (NROW(temp$t_all[!is.na(temp$t_all)]) > 0){
    # there is valid data in the subset, so we can do something
    name<-paste(paste(f, "cdf", sep="-"),"png", sep=".")
    png(name)
    plot(ecdf(temp$t_all))
    dev.off()
  } else {
    # There is no data in the subset, so don't bother
  }
}

N.B. it would have helped had you provided some realistic data and a better description of your data frame. I am guessing from your question that t_all contains the data you wanted to look at.

Upvotes: 0

konvas
konvas

Reputation: 14366

If no observations then the number of rows of temp will be 0, so you can skip to the next value.

if (nrow(temp) == 0) next

Upvotes: 3

Related Questions