[R] ggplot2 applying a function based on facet

[David Winsemius]

On Oct 5, 2009, at 9:39 PM, stephen sefick wrote:

> Look at the bottom of the message for my question
> #here is a little function that I wrote
> USGS <- function(input="discharge", days=7){
> library(chron)
> library(gsubfn)
> #021973269 is the Waynesboro Gauge on the Savannah River Proper (SRS)
> #02102908 is the Flat Creek Gauge (ftbrfcms)
> #02133500 is the Drowning Creek (ftbrbmcm)
> #02341800 is the Upatoi Creek Near Columbus (ftbn)
> #02342500 is the Uchee Creek Near Fort Mitchell (ftbn)
> #02203000 is the Canoochee River Near Claxton (ftst)
>
> a <- "http://waterdata.usgs.gov/nwis/uv?format=rdb&period="
> b <- "&site_no=021973269,02102908,02133500,02341800,02342500,02203000"
> z <- paste(a, days, b, sep="")
> L <- readLines(z)

#trimmed long comment that broke function

> L.USGS <- grep("^USGS", L, value = TRUE)
> DF <- read.table(textConnection(L.USGS), fill = TRUE)
> colnames(DF) <- c("agency", "gauge", "date", "time", "gauge_height",
> "discharge", "precipitation")
> pat <- "^# +USGS +([0-9]+) +(.*)"
> L.DD <- grep(pat, L, value = TRUE)
> library(gsubfn)
> DD <- strapply(L.DD, pat, c, simplify = rbind)
> DDdf <- data.frame(gauge = as.numeric(DD[,1]), gauge_name = DD[,2])
> both <- merge(DF, DDdf, by = "gauge", all.x = TRUE)
>
> dts <- as.character(both[,"date"])
> tms <- as.character(both[,"time"])
> date_time <- as.chron(paste(dts, tms), "%Y-%m-%d %H:%M")
> DF <- data.frame(date_time, both)
> library(ggplot2)
> #discharge
> if(input=="discharge"){
> qplot(as.POSIXct(date_time), discharge, data=DF,
> geom="line")+facet_wrap(~gauge_name,
> scales="free_y")+coord_trans(y="log10")
> }else{
> #precipitation
>
> qplot(as.POSIXct(date_time),
> precipitation, data=subset(DF, precipitation!="NA"),
> geom="line")+facet_wrap(~gauge_name, scales="free_y")
> }
> }
>
> USGS("precip")
>
> I would like to have the cumsum based on the facet gauge_name - in
> other words a cummulative rainfall amount for each gauge_name

You want "the cumsum" of <something> but you have wrapped so much up  
in that function (inlcuding library calls???)  that I cannot see what  
that <something> would be. Not all of us read ggplot calls. The  
canonical route to getting cumsums by a factor is with ave. Something  
like:

  DF$cum_x <- ave(DF$x, DF$fac, cumsum)

-- 
David Winsemius, MD
Heritage Laboratories
West Hartford, CT