Using loops to extract coordinates, match them and write to file - r

Question

I am trying to use for loops (or the apply function as an alternative) to extract coordinates from a data.frame, search for the closest point within the E-OBS gridded dataset, extract the temperature-data for time x1-x2 and write it to another excel file.
While the code works to extract single data points, I seem unable to include this code within a loop and to add the results potentially next to the input-coordinates.
library(sp)
library(raster)
library(ncdf4)
#Coordinates
df
ID site E N
1 1 site_place_date1 7.558758 47.81004
2 2 site_place_date2 7.582749 47.63411
3 3 site_place_date3 7.607968 48.01475
4 4 site_place_date4 7.644660 47.67139
... ... ... ...`
Set coordinates of target point MANUALLY:
lon <- 7.558758 # longitude of location
lat <- 47.81004 # latitude of location
#Mean daily temperature
ncin <- nc_open("tg_0.25deg_reg_v17.0.nc")
print(ncin)
t <- ncvar_get(ncin,"time")
tunits <- ncatt_get(ncin,"time","units")nt <- dim(t)
nt
obsoutput <- ncvar_get(ncin,
start= c(which.min(abs(ncin$dim$longitude$vals - lon)), # look for closest long
which.min(abs(ncin$dim$latitude$vals - lat)), # look for closest lat
1),
count=c(1,1,-1))
DataMeanT <- data.frame(DateN= t, MeanDailyT = obsoutput)
nc_close(ncin)
head(DataMeanT)
#check if there are NAs =999
summary(DataMeanT)
Data = DataMeanT
Data$Date = as.Date(Data$DateN,origin="20000-01-01")
Data$Year = format(Data$Date,"%Y")
Data$Month = format(Data$Date,"%m")
head(Data)
Data$YearMonth = format(Data$Date, format="%Y-%b")
Data_annual = aggregate(("T_AnnualMean" = MeanDailyT) ~ Year,data = Data, FUN = mean,na.action = na.pass)
names(Data_annual)[2] <- "AirT"
head(Data_annual)
#Export table
write.table(Data_annual, "Site_AirTemp.csv", row.names = FALSE, append = FALSE, col.names = TRUE, sep = ", ", quote = TRUE)
The aim is to run the script as part of a loop for all coordinates in df and to write the temperature data to a new data-table with information on site-ID or alternatively into the next columns of df.

Answer 1

Simply wrap your entire process in a defined method and use an apply function to pass in lon/lat coordinates. One great candidate is mapply or its wrapper Map to iterate elementwise between both vectors of df$E and df$N. Also, a third argument, df$site, is passed into method for unique CSV names as right now the same file will be overwritten.
Below some non-assignment lines such as head or summary are removed since they do nothing inside a method. Also context managers, within and with are used to avoid repetition of Data$ for more streamlined data manipulation. The Map call writes to file AND builds a list of aggregated data frames for use later.
Function
my_function <- function(lon, lat, site) {
# Mean daily temperature
ncin <- nc_open("tg_0.25deg_reg_v17.0.nc")
print(ncin)
t <- ncvar_get(ncin,"time")
tunits <- ncatt_get(ncin,"time","units")nt <- dim(t)
# look for closest lon and lat
obsoutput <- ncvar_get(ncin,
start = c(which.min(abs(ncin$dim$longitude$vals - lon)),
which.min(abs(ncin$dim$latitude$vals - lat)),
1),
count = c(1,1,-1))
DataMeanT <- data.frame(DateN = t, MeanDailyT = obsoutput)
nc_close(ncin)
Data <- within(DataMeanT, {
Date <- as.Date(DateN, origin="2000-01-01")
Year <- format(Date,"%Y")
Month <- format(Date,"%m")
YearMonth <- format(Date, format="%Y-%b")
})
Data_annual <- with(Data, aggregate(list("AirT" = MeanDailyT), list(Year=Year),
FUN = mean, na.action = na.pass))
# Export table
write.table(Data_annual, paste0("Site_AirTemp_", site, "_.csv"), row.names=FALSE,
append = FALSE, col.names = TRUE, sep = ", ", quote = TRUE)
# SAVE AGGREGATED DATA FRAME
return(Data_annual)
}
Call
# ITERATE THROUGH EACH LON/LAT PAIR ELEMENTWISE
df_list <- Map(my_function, df$E, df$N, df$site)
# df_list <- mapply(my_function, df$E, df$N, df$site, SIMPLIFY=FALSE) # EQUIVALENT CALL

Answer 2

You can probably do:
library(raster)
b <- brick("tg_0.25deg_reg_v17.0.nc")
e <- extract(b, df[, c('E', 'N')])