我有dataframe dt，如下所示
kmeans sd1 sd2 score gene
B4GALNT1 1 1.138399 0.9302788 0.59238585 B4GALNT1
GATA2 1 1.31817 0.9869005 0.70160114 GATA2
KBTBD8 1 0.2799195 0.25295 2.56658313 KBTBD8
LYPD6 1 0.5885738 0.5277333 1.1797581 LYPD6
MSX1 1 0.2846179 0.5276349 1.31276755 MSX1
NAP1L2 1 0.5778767 0.5252137 1.29646305 NAP1L2
PLA2G4C 1 1.545634 0.3505845 1.02694161 PLA2G4C
SLC6A15 1 3.6862153 1.7656347 0.31940624 SLC6A15
SNORA9 1 49.5847239 23.059789 0.01679016 SNORA9
STX1A 1 4.753248 2.3649298 0.17053974 STX1A
TRNP1 1 54.1230886 19.7797807 0.01907904 TRNP1
AKAP6 2 2.7115279 0.1346139 1.12646609 AKAP6
C1QL3 2 3.1646016 0.3646613 0.78840387 C1QL3
CAMK2N1 2 48.4399203 3.628805 0.05655038 CAMK2N1
CDK5R1 2 3.3858407 0.2249831 0.6292364 CDK5R1
CLSTN2 2 1.0131585 0.162797 1.96050927 CLSTN2
CNTN1 2 3.7191809 0.253088 0.83650197 CNTN1
DGKG 2 0.4607949 0.2333855 1.70445926 DGKG
DPF1 2 1.6369965 0.1873143 1.07265653 DPF1
FAM131A 2 8.7092498 1.763698 0.11250896 FAM131A
我打算使用以下顺序生成下表，方法是使用kmeans列中的组对行进行排名，并根据列kmeans提取每个score组内的排名。所以它应该如下所示
期望的输出：
kmeans sd1 sd2 score gene rank
B4GALNT1 1 1.138399 0.9302788 0.59238585 B4GALNT1 7
GATA2 1 1.31817 0.9869005 0.70160114 GATA2 6
KBTBD8 1 0.2799195 0.25295 2.56658313 KBTBD8 1
LYPD6 1 0.5885738 0.5277333 1.1797581 LYPD6 4
MSX1 1 0.2846179 0.5276349 1.31276755 MSX1 2
NAP1L2 1 0.5778767 0.5252137 1.29646305 NAP1L2 3
PLA2G4C 1 1.545634 0.3505845 1.02694161 PLA2G4C 5
SLC6A15 1 3.6862153 1.7656347 0.31940624 SLC6A15 8
SNORA9 1 49.5847239 23.059789 0.01679016 SNORA9 11
STX1A 1 4.753248 2.3649298 0.17053974 STX1A 9
TRNP1 1 54.1230886 19.7797807 0.01907904 TRNP1 10
AKAP6 2 2.7115279 0.1346139 1.12646609 AKAP6 3
C1QL3 2 3.1646016 0.3646613 0.78840387 C1QL3 6
CAMK2N1 2 48.4399203 3.628805 0.05655038 CAMK2N1 9
CDK5R1 2 3.3858407 0.2249831 0.6292364 CDK5R1 7
CLSTN2 2 1.0131585 0.162797 1.96050927 CLSTN2 1
CNTN1 2 3.7191809 0.253088 0.83650197 CNTN1 5
DGKG 2 0.4607949 0.2333855 1.70445926 DGKG 2
DPF1 2 1.6369965 0.1873143 1.07265653 DPF1 4
FAM131A 2 8.7092498 1.763698 0.11250896 FAM131A 8
但是当我在下面编写代码时，这不是我得到的
dt$rank <- unlist(with(dt, tapply(score, kmeans, function(x) rank(x,ties.method= "first"))))
理想情况下，在每个kemans列中，具有score最高值的行应该被赋予等级1，但这不是我在上面的命令中看到的。我哪里错了？
或者还有其他方法可以达到这个目的吗？
答案 0 :(得分：1)
我们可以使用ave代替tapply执行此操作。 ave的优点是它在获取输出时将保持行的原始顺序
dt$rank <- with(dt, ave(-score, kmeans, FUN = function(x) rank(x, ties.method = "first")))
dt$rank
#[1] 7 6 1 4 2 3 5 8 11 9 10 3 6 9 7 1 5 2 4 8
或使用dplyr
library(dplyr)
dt %>%
group_by(kmeans) %>%
mutate(rank = dense_rank(-score))
dt <- structure(list(kmeans = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), sd1 = c(1.138399,
1.31817, 0.2799195, 0.5885738, 0.2846179, 0.5778767, 1.545634,
3.6862153, 49.5847239, 4.753248, 54.1230886, 2.7115279, 3.1646016,
48.4399203, 3.3858407, 1.0131585, 3.7191809, 0.4607949, 1.6369965,
8.7092498), sd2 = c(0.9302788, 0.9869005, 0.25295, 0.5277333,
0.5276349, 0.5252137, 0.3505845, 1.7656347, 23.059789, 2.3649298,
19.7797807, 0.1346139, 0.3646613, 3.628805, 0.2249831, 0.162797,
0.253088, 0.2333855, 0.1873143, 1.763698), score = c(0.59238585,
0.70160114, 2.56658313, 1.1797581, 1.31276755, 1.29646305, 1.02694161,
0.31940624, 0.01679016, 0.17053974, 0.01907904, 1.12646609, 0.78840387,
0.05655038, 0.6292364, 1.96050927, 0.83650197, 1.70445926, 1.07265653,
0.11250896), gene = c("B4GALNT1", "GATA2", "KBTBD8", "LYPD6",
"MSX1", "NAP1L2", "PLA2G4C", "SLC6A15", "SNORA9", "STX1A", "TRNP1",
"AKAP6", "C1QL3", "CAMK2N1", "CDK5R1", "CLSTN2", "CNTN1", "DGKG",
"DPF1", "FAM131A")), .Names = c("kmeans", "sd1", "sd2", "score",
"gene"), class = "data.frame", row.names = c("B4GALNT1", "GATA2",
"KBTBD8", "LYPD6", "MSX1", "NAP1L2", "PLA2G4C", "SLC6A15", "SNORA9",
"STX1A", "TRNP1", "AKAP6", "C1QL3", "CAMK2N1", "CDK5R1", "CLSTN2",
"CNTN1", "DGKG", "DPF1", "FAM131A"))