Custom grade with dplyr and brilliant
I hope someone can help me with a non-standard estimate when passing variable names to dplyr in a brilliant application. My intention is to be able to select variables to navigate to functions select
and top_n
. I know the function select
has an select_
NSE equivalent , but I'm struggling to work in a brilliant application too.
I have provided an example below which has two lines of comments that I hope to get. The first commented line is for removing the column identified input$var_to_rank
from the output table, and the second commented line (using top_n
) should set the number of top ranked rows to display, and the column that these ranks will be placed on.
library(shiny)
library(dplyr)
data(iris)
shinyApp(
ui = basicPage(
selectInput("species", "species", choices = levels(iris$Species)),
selectInput("var_to_drop", "Variable to drop", choices = names(iris)[3:4]),
selectInput("var_to_rank", "Variable to rank", choices = names(iris)[1:2]),
numericInput("n.obs", "Top N", 5),
tableOutput("table")
),
server = function(input, output) {
output$table <- renderTable({
iris %>%
filter(Species == input$species) %>%
# select_(quote(-input$var_to_drop)) %>%
top_n(5, Sepal.Length)
# top_n(n.obs, input$var_to_rank)
})
}
)
Thanks so much for any help and apologies if this question is answered elsewhere.
source to share
To solve your first problem: this is how you can achieve what you want with NSE
select_(lazyeval::interp(~ -var, var = as.name(input$var_to_drop)))
It might be simpler / shorter, but it works. I know this can be much easier if you want to include rather than remove columns, I cannot define a shorter code that works with-
For your second problem you can achieve the same effect top_n
as like this
cutoff <- iris %>% .[[input$var_to_rank]] %>% sort %>% rev %>% .[input$n.obs]
iris %>% filter_(lazyeval::interp(~ var >= cutoff, var = as.name(input$var_to_rank)))
Just for the sake of completeness, I'm leaving the original answer to the second problem:
For your second problem, this is a solution that works slightly differently. I'm not sure if this is what you want. Usage top_n(5)
could potentially return more than 5 rows, so I do similar but guarantee that only 5 rows are returned
iris %>% arrange_(input$var_to_rank) %>% tail(input$n.obs)
source to share