dplyr：以字符串为列名的条件过滤器

Question

我正在尝试 1) 使用 dplyr::filter 和字符串作为列名和 2) 同时使用 if 语句。然而，虽然其中任何一个本身都可以正常工作，但我很难将它们组合起来，这似乎是由于 if 语句中不能有逗号的问题。下面是 dplyr 0.7.1 的简​​化示例。

欢迎任何见解。非常感谢！！

library("dplyr")

df = data.frame(A=1:6, B=rep(c("good","bad"), 3), C=c("AA","AA","BB","BB","CC","CC"))

# with if statement but not character strings as column names - works
df %>% filter(if(T) {B %in% "good"}) 

# with character string as column names but no if statement - works
df %>% filter_at(vars("B"), any_vars(. %in% "good")) 

# doesn't work when I tried to combine the two
# Error: unexpected ',' in "df %>% filter_at(if(T){vars("B"),"
df %>% filter_at(if(T){vars("B"), any_vars(. %in% "good")})   

# tried to indicate the two parts in the if statement are really one item not two 
# by adding () or {}, and got the same complaint 
df %>% filter_at(if(T){(vars("B"), any_vars(. %in% "good"))}) 
df %>% filter_at(if(T){{vars("B"), any_vars(. %in% "good")}}) 

Answer 1

这行得通：

df %>% filter_at(if(T){vars("B")}, any_vars(. %in% "good"))