问题
How can I store a string (e.g., the column range "cyl:drat, vs:gear"
) so that I can use it in a function where it should not be interpreted as character string?
For example, I would like to execute the following command:
subset(mtcars, select=c(disp:drat, vs:gear))
But assign the content for select
to a variable x
:
x <- as.name("cyl:drat, vs:gear")
subset(mtcars, select=x)
#Error in x[j] : invalid subscript type 'symbol'
library(rlang)
x <- quo(!! sym("cyl:drat, vs:gear"))
subset(mtcars, select=x)
#Error in x[j] : invalid subscript type 'language'
x <- parse_expr("cyl:drat, vs:gear")
subset(mtcars, select=x)
#Error in x[j] : invalid subscript type 'language'
Assigning x <-"cyl"
works, but x <-"cyl:drat"
similarly fails.
Hints as to what format x
should have would already be a welcome start.
回答1:
You missed the c()
in your expression, and you also need to eval
your expressions inside subset
:
library(rlang)
x <- parse_expr("c(cyl:drat, vs:gear)")
subset(mtcars, select=eval(x))
parse_expr
is equivalent to parse
in base R:
x2 = parse(text="c(cyl:drat, vs:gear)")
subset(mtcars, select=eval(x2))
You can also use parse_expr
or parse_exprs
alongside select
from dplyr
, which is where it was intended to be used:
library(dplyr)
select(mtcars, !! x)
or for splicing a list of expressions:
y = parse_exprs("cyl:drat; vs:gear")
select(mtcars, !!! y)
来源:https://stackoverflow.com/questions/47225650/convert-string-to-non-string-input-for-function