Specifying several separate filter conditions and run code for each of the conditions-CodePudding

In my analysis I want to cut and slice my data in several different ways (i.e. want to apply different filters), but otherwise run the same follow-up code (here count) and then ideally glue this together. So what I would do is:

df <- data.frame(x = c(1, 1, 2, 2, 3, 3),
                 y = c(1, 2, 3, 3, 2, 1),
                 z = c(1, 1, 0, 1, 0, 1))

r1 <- df %>%
  filter(x < 3 & y == 3) %>%
  count(z)

r2 <- df %>%
  filter(x != 2 & y == 1) %>%
  count(z)

full_join(r1, r2, by = "z")

which gives:

  z n.x n.y
1 0   1  NA
2 1   1   2

Now, I have a lot of different filter variants, so I'm wondering if there is an easier (tidy) way to achieve this. My idea was that I somehow provide my filter conditions as an input and then apply these filter conditions to my data frame and the bind together the results. I guess this might be able with some purrr magic, but no idea how. Anyone?

CodePudding user response：

Depends in what format do you have your filters, for example you could use eval parse:

filters <- c(
  "x < 3 & y == 3",
  "x != 2 & y == 1"
)

filters %>%
  map(~filter(df, eval(parse(text = .x))) %>% count(z)) %>%
  reduce(full_join, by = "z")

CodePudding user response：

I'm an idiot. I just realized that I could simply do:

df |>
  group_by(z) |> 
  summarize(x = sum(x < 3 & y == 3),
            y = sum(x != 2 & y == 1))