Extract multiple values from a dataset by subsetting with a vector - r

Question

I have a data frame called "Navi", with 72 rows that describe all possible combinations of three variables f,g and h.
head(Navi)
f g h
1 40.00000 80 0.05
2 57.14286 80 0.05
3 74.28571 80 0.05
4 91.42857 80 0.05
5 108.57143 80 0.05
6 125.71429 80 0.05
I have a dataset that also contains these 3 variables f,g and h along with several others.
head(dataset1[,7:14])
# A tibble: 6 x 8
h f g L1 L2 Ref1 Ref2 FR
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 0.02 20 100 53 53 0.501 2.00 2
2 0.02 20 260 67 67 0.200 5.01 5.2
3 0.02 20 420 72 71 0.128 7.83 8.4
4 0.02 20 580 72 72 0.0956 10.5 11.6
5 0.02 20 740 73 73 0.0773 12.9 14.8
6 0.02 20 900 72 71 0.0655 15.3 18
What I'm trying to do is:
for each row in the combinations data frame, filter the dataset by the three variables f,g and h.
Then, if there are exact matches, give me the matching rows of this dataset, then extract the values in the columns "L1" and "FR" in this dataset and calculate the average of them. Save the average value in the vectors "L_M2" and "FR_M2"
If there aren't exact matches, give me the rows where f,g,h in the dataset are closest to f,g,h from the data frame. Then extract all values for L and FR in these rows, and calculate the average. Save the average value in the vectors "L_M2" and "FR_M2".
What I've already tried:
I created two empty vectors where the extracted values shall be saved later on.
Then I am looping over every row of the combinations data frame, filtering the dataset by f,g and h.
The result would be multiple rows, where the values for f,g and h are the same in the dataset as in the row of the combinations data frame.
L_M2 <- vector()
FR_M2 <- vector()
for (i in 1:(nrow(Navi))){
matchingRows[i] <- dataset1[dataset1$P == "input$varP"
& dataset1$Las == input$varLas
& dataset1$Opt == input$varO
& dataset1$f == Navi[i,1]
& dataset1$g == Navi[i,2]
& dataset1$h == Navi[i,3]]
}
The thing is, I don't know what to do from here on. I don't know how to check for rows with closest values by multiple variables, if there are no exact matches...
I only did something more or less similar in the past, but I only checked for the closes "g" value like this:
L_M2 <- vector()
FR_M2 <- vector()
for (i in 1:(nrow(Navi))){
matchingRows[i] <- dataset1[dataset1$P == "input$varP"
& dataset1$Las == input$varLas
& dataset1$Opt == input$varO
& dataset1$f == Navi[i,1]
& dataset1$g == Navi[i,2]
& dataset1$h == Navi[i,3]]
for (i in 1:(nrow(Navi)){
Differences <- abs(Navi[i,2]- matchingRows$G)
indexofMin <- which(Differences == min (Differences))
L_M2 <- append(L_M2, matchingRows$L[[indexofMin]], after = length(L_M2))
FR_M2 <- append(FR_M2, matchingRows$FR[[indexofMin]], after = length(FR_M2))
}
So can anybody tell me how to achieve this extraction process?I am still pretty new to R, so please tell me If I made a rookie mistake or forgot to include some crucial information. Thank you!

Answer 1

First convert your data into dataframe (if not done before).
Navi <- data.frame(Navi)
Savi <- data.frame(dataset1[,7:14])
Then use merge to filter your lines:
df1 <- merge(Navi, Savi, by = c("f","g","h"))
Save "L1" and "FR" average from df1:
Average1 <- ((df1$L1+df1$FR)/2)
Get you your new Navi dataframe which doen not have exact match on f,g,h columns
Navi_new <- Navi[!duplicated(rbind(df1, Navi))[-seq_len(nrow(df1))], ]
For comparing the values with nearest match:
A1 <- vapply(Navi_new$f, function(x) x-Savi$f, numeric(3))
A2 <- apply(abs(A1), 2, which.min)
B1 <- vapply(A1$g, function(x) x-Savi$g, numeric(3))
B2 <- apply(abs(B1), 2, which.min)
C1 <- vapply(B1$g, function(x) x-Savi$g, numeric(3))
C2 <- apply(abs(C1), 2, which.min)
You can use C2 dataframe to get the average of "L1" and "FR" like 3 steps back.