How to grab selected event data from a ggplotly chart? - r

Question

I'm aware of https://plot.ly/r/shinyapp-plotly-events/ and have been using it as a guide. But the ggplot element I'm converting to plotly is the output from the fviz_dend function of the factoextra package. Here's a minimum shiny app example I'm working with:
library(factoextra)
library(plotly)
library(shiny)
library(DT)
ui <- fluidPage(
plotlyOutput("ggp"),
verbatimTextOutput("selected_points"),
DT::dataTableOutput("filtered_table")
)
server <- function(input, output, session) {
## ggplot output
fviz <- fviz_dend(
x = hclust(dist(mtcars)),
k = 5,
show_labels = TRUE,
type = "phylogenic",
phylo_layout = "layout_as_tree",
color_labels_by_k = TRUE,
palette = "igv"
)
## convert to ggplotly
ggfviz <- ggplotly(fviz)
## add keys
for (i in seq(7, 11)) {
ggfviz[["x"]][["data"]][[i-5]][["key"]] <-
as.character(ggfviz[["x"]][["data"]][[i]][["text"]])
}
output$ggp <- renderPlotly({
ggfviz
})
output$selected_points <- renderPrint({
event_data("plotly_selected")[5]
})
output$filtered_table <- DT::renderDataTable(
mtcars[which(rownames(mtcars) == event_data("plotly_selected")[5]), ],
)
}
shinyApp(ui, server)
So I'm trying to use the key accessed with event_data("plotly_selected")[5] in order to filter the data table, and while event_data("plotly_selected")[5] does show the key per output$selected_points, it is somehow not passed to the datatable filter.

Answer 1

It looks like event_data will return a data frame with multiple rows. Instead of filtering with == you will need %in% instead to see which multiple cars are contained within the multiple possible selections from plotly_selected. In addition, even though you subset by column 5, you still have a data frame, and need to include the column key only for filtering (containing a vector of cars). This should work:
mtcars[which(rownames(mtcars) %in% event_data("plotly_selected")$key), ]
Or
mtcars[which(rownames(mtcars) %in% event_data("plotly_selected")[["key"]]), ]