我正在尝试在R:在某些条件下在向量中提取最大值,但我继续遇到错误
Error in list(id.2 = c(3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, :
invalid subscript type 'integer'
代码如下:
require(dplyr)
dat <- read.table(header = TRUE, text = "id name year job job2 cumu_job2
1 Jane 1980 Worker 0 0
1 Jane 1981 Manager 1 1
1 Jane 1982 Sales 0 0
1 Jane 1983 Sales 0 0
1 Jane 1984 Manager 1 1
1 Jane 1985 Manager 1 2
1 Jane 1986 Boss 0 0
2 Bob 1985 Worker 0 0
2 Bob 1986 Sales 0 0
2 Bob 1987 Manager 1 1
2 Bob 1988 Manager 1 2
2 Bob 1989 Boss 0 0
3 Jill 1989 Worker 0 0
3 Jill 1990 Boss 0 0")
dat %.%
group_by(id) %.%
mutate(
all_jobs = sum(unique(job) %in% c("Sales","Manager","Boss")),
cumu_max = max(cumu_job2)
) %.%
filter(all_jobs == 3, job %in% c("Sales","Boss"))
Source: local data frame [5 x 8]
Groups: id
id name year job job2 cumu_job2 all_jobs cumu_max
1 1 Jane 1982 Sales 0 0 3 2
2 1 Jane 1983 Sales 0 0 3 2
3 1 Jane 1986 Boss 0 0 3 2
4 2 Bob 1986 Sales 0 0 3 2
5 2 Bob 1989 Boss 0 0 3 2
示例代码也对我有用。但是我发现,如果尝试此尝试,我可以将类似的错误重复出现:
dat %.%
group_by(dat$id) %.%
mutate(
all_jobs = sum(unique(job) %in% c("Sales","Manager","Boss")),
cumu_max = max(cumu_job2)
) %.%
filter(all_jobs == 3, job %in% c("Sales","Boss"))
也就是说,如果我键入" group_by(dat $ id)"而不是" group_by(id)"
bug
样本代码也对我有用。但是,正如Schnee提到的那样,您可以通过group_by(dat $ id)替换group_by(id)来创建类似的错误。可再现的代码:
dat1 <- data.frame(x=c('A','A','B','B'), y=c('A','B','C','D'), val = 1:4)
dat2 <- data.frame(val = 1:4)
dat_group <- data.frame(x=c('A','A','B','B'))
# invalid subscript type 'integer'
dat1 %>%
group_by(dat1$x) %>%
mutate(y = sum(unique(y) %in% c("A","B","C")))
# invalid subscript type 'list'
dat2 %>%
group_by(dat_group$x) %>%
mutate(y = sum(unique(y) %in% c("A","B","C")))
虽然第一个通常只是错字(您可以用x替换dat $ x),但第二个可能是有效的用例(尽管我建议加入以使其更清洁)。
解决方案
dplyr软件包不喜欢" $"的用法。尝试使用'[',例如:
dat1[,'x']
引用变量也有效:
dat1$'x'
完整代码:
dat1 %>%
group_by(dat1[,'x']) %>%
mutate(y = sum(unique(y) %in% c("A","B","C")))
dat1 %>%
group_by(dat1$'x') %>%
mutate(y = sum(unique(y) %in% c("A","B","C")))
另请参阅https://github.com/hadley/dplyr/issues/433或https://github.com/hadley/dplyr/issues/1554