Skip to contents

If pickable == TRUE, tooltip defines which columns (and their order) will be displayed when an object is hovered. The tooltip will be displayed as a transposed table (1 row per column) and is styled according to the rdeck() theme. The tooltip layout is inspired by kepler.gl.

Tooltips support tidy-select expressions (renaming isn't supported and will fail silently) and boolean expressions. Tidy-select support is limited when layer data is a URL or NULL.

Similar to dplyr::select(), names are interpreted as column / field names. Use the injection operators to pass the value of a name from the environment, e.g. !!my_column.

Always supported

Tooltips support different arguments depending on the value of the layer data, but the following arguments are always supported:

  • NULL | NA | FALSE -> no tooltip

  • TRUE -> all columns. This will be translated to tidyselect::everything() when data is a data frame or a tile json object

  • character vector of column names, e.g. c("foo", "bar")

  • a c() expression of bare names, e.g. c(foo, bar)

  • a bare name, e.g. my_col

Data is a data.frame or tile_json object

When layer data is a data.frame() or tile_json() object, tooltip supports tidy-select, including selection helpers. The following are some examples of supported expressions:

Like dplyr::select(), referencing non-existent columns results in an error. Use tidyselect::any_of() to include columns that may not exist.

All geometry columns will be forcibly removed from the tooltip.

Data is a string or NULL

There is no column / field name validation when When layer data is a string (interpreted as a URL) or NULL. Including non-existent names in the tooltip will not cause errors in the client; these columns will be given a value of null in the rendered tooltip.