问题
This is a continuation of a question that I had originally posted at http://r.789695.n4.nabble.com/subset-between-data-table-list-and-single-data-table-object-tp4673202.html . Matthew had suggested that I post my question here so I am doing that now.
This is my input below:
library(data.table)
library(pracma) # for the interp1 function
tempbigdata1 <- data.table(c(14.80, 14.81, 14.82), c(7900, 7920, 7930), c("02437100", "02437100", "02437100"))
tempbigdata2 <- data.table(c(9.98, 9.99, 10.00), c(816, 819, 821), c("02446500", "02446500", "02446500"))
tempbigdata3 <- data.table(c(75.65, 75.66, 75.67), c(23600, 23700, 23800), c("02467000", "02467000", "02467000"))
tempsbigdata <- rbind(tempbigdata1, tempbigdata2, tempbigdata3)
setnames(tempsbigdata,c("y", "x", "site_no"))
setkey(tempsbigdata, site_no)
tempsbigdata
y x site_no
1: 14.80 7900 02437100
2: 14.81 7920 02437100
3: 14.82 7930 02437100
4: 9.98 816 02446500
5: 9.99 819 02446500
6: 10.00 821 02446500
7: 75.65 23600 02467000
8: 75.66 23700 02467000
9: 75.67 23800 02467000
aimsmall <- data.table(c("02437100", "02446500", "02467000"), c(3882.65, 819.82, 23742.37), c(1830.0, 382.0, 10400.0))
setnames(aimsmall,c("site_no", "mean", "p50"))
setkey(aimsmall, site_no)
aimsmall
site_no mean p50
1: 02437100 3882.65 1830
2: 02446500 819.82 382
3: 02467000 23742.37 10400
I am using this code to generate the interpolated tempsbigdata$y using the aimsmall$mean values by the site_no:
meanpre <- tempsbigdata[,if(aimsmall$mean > min(tempsbigdata$x){
interp1(tempsbigdata$x, tempsbigdata$y,
xi = aimsmall$mean, method ="linear")},by=site_no]
This is the output from the function meanpre, but it is not correct.
meanpre
site_no V1
1: 02437100 12.07599
2: 02437100 9.99410
3: 02437100 19.56813
4: 02446500 12.07599
5: 02446500 9.99410
6: 02446500 19.56813
7: 02467000 12.07599
8: 02467000 9.99410
9: 02467000 19.56813
This is what I would like to get:
meanpre
site_no V1
1: 02446500 9.99
2: 02467000 75.66
Any suggestions? Thank you.
UPDATE 1:
Hugh, I used the approx function in the past and it is not accurate for my data; however, the interp1 function in pracma is accurate. The mean and p50 columns in aimsmall & the x values in tempsbigdata are discharge values. The y in tempsbigdata represent gage heights. I am using the interp1 function to determine the appropriate gage height or y value for the discharge values or mean (and p50).
Frank, thank you for your advice and suggested code. This is the output for your suggested code:
tempsbigdata[aimsmall][,if(mean[1] > min(x)){interp1(tempsbigdata$x,tempsbigdata$y, xi = aimsmall$mean, method ="linear")},by=site_no]
site_no V1
1: 02446500 12.07599
2: 02446500 9.99410
3: 02446500 75.66424
4: 02467000 12.07599
5: 02467000 9.99410
6: 02467000 75.66424
When I run the following code I get the result below:
interp1(tempsbigdata$x, tempsbigdata$y, xi = aimsmall$mean, method ="linear")
[1] 12.07599 9.99410 75.66424
Is there any way to get this in return? Thank you.
site_no V1
1: 02446500 9.99
2: 02467000 75.66
UPDATE 2
Frank, thank you and I have added the code to make it easier to have the data in R. Pracma is an R package of numerical method routines that were ported from GNU Octave [similar to MATLAB(R)] to R. The interp1 function is a one-dimensional interpolation function.
Frank, that was perfect (your last comment about the R code for "do stuff"):
tempsbigdata[aimsmall][,if(mean[1] > min(x)){interp1(x, y, xi = mean[1], method ="linear")},by=site_no]
site_no V1
1: 02446500 9.99410
2: 02467000 75.66424
来源：https://stackoverflow.com/questions/18165373/interpolation-of-grouped-data-using-data-table