Selecting an entire row of a minimum element in a dataframe

Question

I am new to R and I am trying to get the minimum Distance value and the corresponding "Record2_ID" value for every unique "Record1_ID" value for the below dataframe

Record1_ID  Record2_ID  Distance
6       10_Bil      0.95337476
6       11_Bla      0.852558044
6       12_Bon      1
6       13_Bra      1
684     78_Lip      0.957437173
684     79_Lip      1
684     80_Liv      0.950852681
684     81_Lun      0.914874347
3065        136_Pri     1
3065        137_Pro     0.895742793
3065        138_Rec     0.895742793
3065        139_Ren     0.934061953

I used the function tapply(x$Distance_Cosine, cosine_dist_type_data$Record1_rowID, min), but using tapply I am not getting "Record2_rowID" values. Ideally the output should be

Record1_ID  Record2_ID  Min_Distance
6       11_Bla      0.852558044
684     81_Lun      0.914874347
3065        137_Pro     0.895742793

Can this be done using sapply or any other function. Thanks for the help

Answer 1

Or you may use the base function ave

df[df$Distance == ave(df$Distance, df$Record1_ID, FUN = min), ]
#    Record1_ID Record2_ID  Distance
# 2           6     11_Bla 0.8525580
# 8         684     81_Lun 0.9148743
# 10       3065    137_Pro 0.8957428
# 11       3065    138_Rec 0.8957428

Answer 2

library(data.table)
df = data.table(read.table(header = T, text = "
            Record1_ID  Record2_ID  Distance
6       10_Bil      0.95337476
6       11_Bla      0.852558044
6       12_Bon      1
6       13_Bra      1
684     78_Lip      0.957437173
684     79_Lip      1
684     80_Liv      0.950852681
684     81_Lun      0.914874347
3065        136_Pri     1
3065        137_Pro     0.895742793
3065        138_Rec     0.895742793
3065        139_Ren     0.934061953
            "))

df[, Min_Distance := min(Distance), by = Record1_ID]
df[Distance == Min_Distance,]

---

Or slightly more straightforward:

df[, .SD[Distance == min(Distance)], by=Record1_ID]

.SD contains the Subset of Data for each group. We just select the rows we want on that subset corresponding to min(Distance).

Answer 3

If that's a dataframe, you want to look at plyr, specifically the ddply function. Not tremendously elegant, but try...

min_vals.df <- ddply(.data = df,
                     .variables = "Record1_ID",
                     .fun = function(x){
                         return(x[x$Distance == min(x$Distance),c("Record2_ID","Distance")])

Plyr and it's successor, dplyr, are "apply for data frames", iterating through each unique permutation of .variables and performing whatever function you want on the resulting data.

Answer 4

or without plyr:

blah <- lapply(split(df, df["Record1_ID"]), function(x) x[which.min(x$Distance),])
min_vals.df <- do.call(rbind, blah)

(edit) a modification to include all min. values (if there are ties)

blah <- lapply(split(df, df["Record1_ID"]), function(x) subset(x, Distance==min(Distance)))
min_vals.df <- do.call(rbind, blah)

Answer 5

Or with dplyr:

require(dplyr)

df %.% group_by(Record1_ID) %.% filter(Distance == min(Distance))