Issue when your data are too good

[kemarshall]

Hi,

I found this a really useful set of functions and I'm really excited to give them a whirl. But there is a bit of an issue if your data are too good--if you're not removing any data then you actually end up with an empty dataframe. Try getInterpolatedDataByCity("Vancouver, Canada",station.list,10,2015,2016,100,6,0.3) for an example.

So I managed to figure out where the bugs that cause this are. The first is in the combineWeatherDFs function where trackNull <- which(names(dfList$station_data)=="" | names(dfList$station_data) %in% NA) returns a logical(0) if you have data without any issues, which ends up removing all your data in the next line. I kludged through it by adding trackNull <- c(trackNull,1000) as a next line here, but I'm sure there's a more elegant solution for that.

The next bug is in the filterStationData function where rm.df <- data.frame(table(substr(rm.all, 1, nchar(rm.all)-7))) returns a null frame if there is nothing to substring out. I fixed it by adding this next if(dim(rm.df)[1] == 0) { rm.df <- kept.df }. Again, not the tidiest but it seemed to work.

Anyway, I hope this helps and I'm super excited to see this package in more use!

Cheers,
Katie

[mpiccirilli]

You can make a pull request with the fixes if you'd like

[kemarshall]

I'm not super confident with git, and my fixes are kind of ugly. But happy to try it out this weekend if it would be helpful!

[mpiccirilli]

No problem. I'm always open to helpful contributions!