How to programmatically provide a list of filters to apply via dplyr and filter_

Question

I'd like to create a list of filters to apply to a dataframe. Something like:

filters = list(cyl=4, am=1)

and then apply that to the 'mtcars' dataframe, to get just the records with cyl=4 and am=1. I can do this:

filter_(mtcars, 
        lazyeval::interp(~ val == var, val = as.name(names(filters[1])), 
                                       var = filters[[1]]))

But that only pick up the 1st entry in the filters list.

What is the idiomatic way to apply all of the filters?

(I'm trying to create a somewhat generic function that can accept a dataframe and a criteria set and will output transformations. Right now, equality is fine for the criteria, but more general idioms would be nice)

Answer 1

Define filter as

filter1 <- ~(cyl==4 & am==1)

or

filter1 <- "cyl==4 & am==1"

or

library(lazyeval)
filter1 <- lazy(cyl==4 & am==1)

and use as

mtcars %>% filter_(filter1)

Example with a function:

get_cars_with_filter <- function(my_filter) {
  mtcars %>% filter_(lazy(my_filter))
}

get_cars_with_filter(cyl==4 & am == 1)

Answer 2

All answers here are good but if you strictly want to keep your original way of defining the filter (via a list) you can simply convert it to a string an pass it to filter_ like

filter_(mtcars, paste(names(filters), filters, sep = "==", collapse = "&"))

(with the added benefit that this allows flexibility in terms of the logical operators you use, e.g. can change collapse to "|" etc)