R 粘贴三列长度不等的列

假设我有下面的数据框，并希望将这三列合并到第四列中

df = data.frame(one=c("honda","ford","kia","bmw"),
                two=c("subaru","","","lexus"),
                three=c("","","","ford"))

我可以用浆糊把这三个结合起来

df$label = paste(df$one, df$two, df$three, sep=" - ")
df

不幸的是，由于数据帧中的“空”值，这会导致以下情况

> df
    one    two three              label
1 honda subaru        honda - subaru - 
2  ford                      ford -  - 
3   kia                       kia -  - 
4   bmw  lexus  ford bmw - lexus - ford

我可以只清理标签列并删除滞后的破折号，但有没有办法在粘贴中这样做，以便在两个或三个值中没有一个值时，不使用额外的破折号

应该是这样的：

> df
    one    two three              label
1 honda subaru            honda - subaru
2  ford                         ford
3   kia                          kia
4   bmw  lexus  ford    bmw - lexus - ford

> df$label <- apply(df, 1, function(x) paste(x[x!=''], collapse=' - '))
> df
    one    two three              label
1 honda subaru           honda - subaru
2  ford                            ford
3   kia                             kia
4   bmw  lexus  ford bmw - lexus - ford