Difference between airDF[is_vacation, airDF$ArrDelay] (where is_vacation is a logical vector) and ai-CodePudding

What is the difference between (is_vacation is a logical vector):

tapply(X = airDF[is_vacation, airDF$ArrDelay], airDF[is_vacation, airDF$FromTo], mean, na.rm = TRUE)

tapply(X = airDF[is_vacation, 'ArrDelay'], airDF[is_vacation, 'FromTo'], mean, na.rm = TRUE)

In the first option i got error:

Error in [.data.frame(airDF, is_vacation, airDF$FromTo) : no columns were selected.

CodePudding user response：

The first option is wrong, as instead of the column names or index, it is the column values used in the j where as second option is fine

 tapply(mtcars[1:5, 'mpg'], mtcars[1:5, 'gear'], mean)
    3     4 
20.05 21.60