Сохранять строки в списке data.frames в соответствии с условиями

Question

У меня есть данные data.frame следующим образом.

ind <- c("22", "58", "57", "43", "23", "90", "45", "21", "27", "67", 
     "17", "50", "71", "49", "92", "36", "62", "83", "02", "95") 

gp <- c("G1", "G1", "G1", "G23", "G23", "G28", "G28", "G29", "G53", 
     "G56", "G56", "G67", "G77", "G77", "G79", "G79", "G79", 
     "G82", "G82", "G82") 
id <- c("T297170", "T304934", "T437551", "T572358", "T572359", "T118839", 
     "T304962", "T594651", "T113085", "T304969", "T444487", "T296315", 
     "T305008", "T437400", "T113089", "T305032", "T557004", "T445004", 
     "T445004", "T78642") 
nm <- c("MaskedMarvel", "DecemberSnowflakes", "MaskedMarvel", 
     "WarmPuppy", "WarmPuppy", "SpringDance", "SpringDance", 
     "RoastedMarshmallows", "TrickorTreat", "FrisbeeSailing", 
     "FrisbeeSailing", "GreatPumpkin", "PumpkinHelmet", "PumpkinHelmet", 
     "GoodSport", "GoodSport", "GoodSport", "GiftGettingSeason", 
     "EasterBeagle", "EasterBeagle") 

data <- data.frame(ind, id, gp, nm) 
data 
    ind  id gp     nm 
1 22 T297170 G1  MaskedMarvel 
2 58 T304934 G1 DecemberSnowflakes 
3 57 T437551 G1  MaskedMarvel 
4 43 T572358 G23   WarmPuppy 
5 23 T572359 G23   WarmPuppy 
6 90 T118839 G28   SpringDance 
7 45 T304962 G28   SpringDance 
8 21 T594651 G29 RoastedMarshmallows 
9 27 T113085 G53  TrickorTreat 
10 67 T304969 G56  FrisbeeSailing 
11 17 T444487 G56  FrisbeeSailing 
12 50 T296315 G67  GreatPumpkin 
13 71 T305008 G77  PumpkinHelmet 
14 49 T437400 G77  PumpkinHelmet 
15 92 T113089 G79   GoodSport 
16 36 T305032 G79   GoodSport 
17 62 T557004 G79   GoodSport 
18 83 T445004 G82 GiftGettingSeason 
19 02 T445004 G82  EasterBeagle 
20 95 T78642 G82  EasterBeagle 

Я хочу сохранить только те строки, в которых группы (указанные в gp имеют разнородные элементы в столбце nm. Группы с отдельными элементами в nm не должны быть сохранены.

Я получаю желаемый результат с код ниже.

# Split the data.frame into a list of data.frames by groups 
data <- lapply(levels(data$gp),function(i) data[which(data$gp==i),]) 
# Remove groups with single elements in data$nm 
data <- data[!sapply(data, function(i) nrow(i) == 1)] 
# Remove groups with multiples of only a single element in data$nm 
data <- data[!sapply(seq_along(data), function(i) length(unique(data[[i]][,4])) == 1)] 
# cbind the data.frames  
data <- do.call(rbind, data) 
# fix the levels in factor variables 
data <- droplevels(data) 
data 
    ind  id gp     nm rm 
1 22 T297170 G1  MaskedMarvel TRUE 
2 58 T304934 G1 DecemberSnowflakes FALSE 
3 57 T437551 G1  MaskedMarvel TRUE 
18 83 T445004 G82 GiftGettingSeason FALSE 
19 02 T445004 G82  EasterBeagle TRUE 
20 95 T78642 G82  EasterBeagle TRUE 

есть ли более элегантный способ для достижения этого результата, может быть, в одном шаге в базе R?

Answer 1

Вы можете попробовать

data1 <- data[with(data, ave(as.character(nm), gp, FUN= 
          function(x) length(unique(x)))>1),] 
transform(data1, rm=ave(as.character(nm), 
    gp, FUN=function(x) duplicated(x)|duplicated(x,fromLast=TRUE))) 
# ind  id gp     nm rm 
#1 22 T297170 G1  MaskedMarvel TRUE 
#2 58 T304934 G1 DecemberSnowflakes FALSE 
#3 57 T437551 G1  MaskedMarvel TRUE 
#18 83 T445004 G82 GiftGettingSeason FALSE 
#19 02 T445004 G82  EasterBeagle TRUE 
#20 95 T78642 G82  EasterBeagle TRUE 

Или с помощью data.table

setDT(data)[,.SD[length(unique(nm))>1], by=gp][, 
     rm:= duplicated(nm)|duplicated(nm,fromLast=TRUE) ,by=gp][] 
# gp ind  id     nm rm 
#1: G1 22 T297170  MaskedMarvel TRUE 
#2: G1 58 T304934 DecemberSnowflakes FALSE 
#3: G1 57 T437551  MaskedMarvel TRUE 
#4: G82 83 T445004 GiftGettingSeason FALSE 
#5: G82 02 T445004  EasterBeagle TRUE 
#6: G82 95 T78642  EasterBeagle TRUE 

Или с помощью dplyr

library(dplyr) 
data %>% 
     group_by(gp) %>% 
     filter(n_distinct(nm)>1) %>% 
     mutate(rm=duplicated(nm)|duplicated(nm,fromLast=TRUE))