I am trying to add sparklines to multiple columns of a Data Table in R. I am able to get a single Sparkline column to render correctly, but when I try to render a sparkline in a second column based on different underlying data than the first, the same sparkline is displayed in the second column. I imagine there is some snippet of code needed in the fnDrawCalback option, but I have not been able to find much out there addressing my issue.
Here is a reproducible example for both a single sparkline (works) and a two sparklines (just repeats the first sparkline in both columns)
# create data with single sparkline column
library(sparkline)
library(DT)
spark_data <- data.frame(
id = c('spark1', 'spark2'),
spark = c(
spk_chr(values = 1:3, elementId = 'spark1'),
spk_chr(values = 3:1, elementId = 'spark2')
)
)
tbl <- datatable(spark_data, escape = F,
rownames = F
, options = list(fnDrawCallback = htmlwidgets::JS('function(){
HTMLWidgets.staticRender();
}'))
)
spk_add_deps(tbl)
# create data with two sparkline columns
library(sparkline)
library(DT)
spark_data <- data.frame(
id = c('spark1', 'spark2'),
spark = c(
spk_chr(values = 1:3, elementId = 'spark1'),
spk_chr(values = 3:1, elementId = 'spark2')
),
second_spark = c(spk_chr(values = 3:1, elementId = 'spark1'),
spk_chr(values = 1:3, elementId = 'spark2'))
)
tbl <- datatable(spark_data, escape = F,
rownames = F
, options = list(fnDrawCallback = htmlwidgets::JS('function(){
HTMLWidgets.staticRender();
}'))
)
spk_add_deps(tbl)
I created a datatable in a flexdashboard with a checkbox, but the checkbox flows off the page. I tried to adjust the padding {data-padding = 10} but nothing changed. Below is the code and a picture of what the dashboard looks like. How do I move everything to the right so that it's aligned with the title of the page?
---
title: "School Dashboard"
author: "Shannon Coulter"
output:
flexdashboard::flex_dashboard:
orientation: rows
social: menu
source_code: embed
theme: spacelab
---
```{r}
library(tidyverse)
library(crosstalk)
library(DT)
library(flexdashboard)
```
Student Lookup
================================================================================
### Chronic Absenteeism Lookup
```{r ca-lookup, echo=FALSE, message=FALSE, warning=FALSE}
ican_tab <- tibble(
year = c("2022", "2022", "2022", "2022", "2022"),
date = c("March", "March","March","March","March"),
school = c("ABC", "CDE","ABC","DEF","GHI"),
grade = c("6th", "7th","8th","4th","5th"),
race_eth = c("White", "Hispanic","White","Filipino","White"),
abs_levels = c("Not At-Risk of Chronic Absenteeism", "At-Risk of Chronic Absenteeism",
"Severe Chronic Absenteeism", "Severe Chronic Absenteeism",
"Moderate Chronic Absenteeism")
)
sd <- SharedData$new(ican_tab)
bscols(list(
filter_checkbox("abs_levels", "Level", sd, ~ abs_levels, inline = TRUE),
datatable(
sd,
extensions = c("Buttons",
"Scroller"),
options = list(
autoWidth = TRUE,
scrollY = F,
columnDefs = list(list(
className = 'dt-center',
targets = c(2, 3, 4, 5)
)),
lengthMenu = c(5, 10, 25, 100),
dom = "Blrtip",
deferRender = TRUE,
scrollY = 300,
scroller = TRUE,
buttons = list('copy',
'csv',
'pdf',
'print')
),
filter = "top",
style = "bootstrap",
class = "compact",
width = "100%",
colnames = c(
"Year",
"Date",
"School",
"Grade",
"Race",
"Level"
)
) %>%
formatStyle('abs_levels',
backgroundColor = styleEqual(
unique(ican_tab$abs_levels),
c(
"#73D055ff",
"#95D840FF",
"#B8DE29FF",
"#DCE319FF"
)
))
))
```
[![enter image description here][1]][1]
The easiest way to address this is probably to add style tags to your dashboard. You can put this anywhere. I usually put it right after the YAML or right after my first R chunk, where I just place my knitr options and libraries. This does not go inside an R chunk.
<style>
body { /*push content away from far right and left edges*/
margin-right: 2%;
margin-left: 2%;
}
</style>
Update based on your updated question and comments
I don't have the content around your table, so I will give you a few options that work. For the most part, any one option won't be enough. You can mix and match the options that work best for you.
This is what I've got for the original table:
Option 1: you can use CSS to push the table away from the edges (as in my original response
Option 2: change the font sizes
Option 3: constrain the size of the datatable htmlwidget
Option 4: manually make the columns narrower
Option 5: alter the filter labels (while keeping the same filters and data)
Aesthetically looks the best? It depends on what else is on the dashboard.
I think you will need the original CSS (option 1, in my original answer) regardless of what other options you choose to use.
Option 1 is above
Option 2
To change the font sizes, you have to modify the filter_checkbox and the datatable after they're made. Instead of presenting all of the programming code, I'm going to show you want to add or modify and how I broke down the objects.
Your original code for filter_checkbox remains the same. However, you'll assign it to an object, instead of including it in bscols.
Most of the code in your datatable will remain the same. there is an addition to the parameter options. I've included the original and change for that parameter.
# filter checkbox object
fc = filter_checkbox(...parameters unchanged...)
fc$attribs$style <- css(font.size = "90%") # <-change the font size
dt = datatable(
...
...
options = list( # this will be modified
autoWidth = TRUE, # <- same
scrollY = F, # <- same
initComplete = JS( # <- I'M NEW! change size of all font
"function(settings, json) {",
"$(this.api().table().container()).css({'font-size': '90%'});",
"}"),
columnDefs = list( # <- same
list(className = 'dt-center', targets = c(2, 3, 4, 5))),
...
... # remainder of datatable and formatStyles() original code
)
# now call them together
bscols(list(fc, dt))
The top version is with 90% font size, whereas the bottom is the original table.
Option 3
To constrain the size of the datatable widget, you'll need to create the object outside of bscols, like I did in option 2. If you were to name your widget dt as in my example, this is how you could constrain the widget size. This example sets the datatable to be 50% of the width and height viewer screen (or 1/4 of the webpage). Keep in mind that the filters are not part of the widget, so in all, the table is still more than 1/4th of the webpage. You will have to adjust the size for your purposes, of course. I recommend using a dynamic sizing mechanism like vw, em, rem, and the like.
dt$sizingPolicy$defaultWidth <- "50vw"
dt$sizingPolicy$defaultHeight <- "40vh"
The top image has options 1, 2, and 3; the bottom is the original table.
Option 4
To modify the width of the columns, you can add this modification to the parameter options in you call to datatable. This could be good, because most of the columns don't require as much width as the last column. However, if you change the font size or scale the table, it will change the font size dynamically, so this option may not be necessary.
Despite using em here, in the course of this going from R code to an html_document, it was changed to pixels. So this is not dynamically sized. (Not a great idea! Sigh!)
columnDefs = list(
list(className = 'dt-center', targets = c(2, 3, 4, 5)),
list(width = '5em', targets = c(1,2,3,4,5))), # <- I'm NEW!
Option 5
For this option, I took the programming behind crosstalk::filter_checkbox() and modified the code a bit. I changed the function to filter_checkbox2(). If you use it, you can render it both ways and just keep the one you like better.
This first bit of code is the three functions that work together to create a filter_checkbox object with my modifications so that you can have a label that isn't exactly the same as the levels.
It's important to note that the filters are alphabetized by datatable. It doesn't matter if they're factors, ordered, etc. If you use this new parameter groupLabels, they need to be in an order that aligns with the levels when they're alphabetized.
I put this code in an include=F chunk by itself:
# this is nearly identical to the original function
filter_checkbox2 = function (id, label, sharedData, group,
groupLabels = NULL, # they're optional
allLevels = FALSE, inline = FALSE, columns = 1) {
options <- makeGroupOptions(sharedData, group,
groupLabels, allLevels) # added groupLabels
labels <- options$items$label
values <- options$items$value
options$items <- NULL
makeCheckbox <- if (inline)
inlineCheckbox
else blockCheckbox
htmltools::browsable(attachDependencies(tags$div(id = id,
class = "form-group crosstalk-input-checkboxgroup crosstalk-input",
tags$label(class = "control-label", `for` = id, label),
tags$div(class = "crosstalk-options-group",
crosstalk:::columnize(columns,
mapply(labels, values, FUN = function(label, value) {
makeCheckbox(id, value, label)
}, SIMPLIFY = FALSE, USE.NAMES = FALSE))),
tags$script(type = "application/json", `data-for` = id,
jsonlite::toJSON(options, dataframe = "columns",
pretty = TRUE))),
c(list(crosstalk:::jqueryLib()),crosstalk:::crosstalkLibs())))
}
inlineCheckbox = function (id, value, label) { # unchanged
tags$label(class = "checkbox-inline",
tags$input(type = "checkbox",
name = id, value = value),
tags$span(label))
}
# added groupLabels (optional)
makeGroupOptions = function (sharedData, group, groupLabels = NULL, allLevels) {
df <- sharedData$data(withSelection = FALSE, withFilter = FALSE,
withKey = TRUE)
if (inherits(group, "formula"))
group <- lazyeval::f_eval(group, df)
if (length(group) < 1) {
stop("Can't form options with zero-length group vector")
}
lvls <- if (is.factor(group)) {
if (allLevels) {levels(group) }
else { levels(droplevels(group)) }
}
else { sort(unique(group)) }
matches <- match(group, lvls)
vals <- lapply(1:length(lvls), function(i) {
df$key_[which(matches == i)]
})
lvls_str <- as.character(lvls)
if(is.null(groupLabels)){groupLabels = lvls_str} # if none provided
if(length(groupLabels) != length(lvls_str)){ # if the # labels != the # groups
message("Warning: The number of group labels does not match the number of groups.\nGroups were used as labels.")
groupLabels = lvls_str
}
options <- list(items = data.frame(value = lvls_str, label = groupLabels, # changed from lvls_str
stringsAsFactors = FALSE), map = setNames(vals, lvls_str),
group = sharedData$groupName())
options
}
When I used this new version of I changed label = "Level" to label = "Chronic Absenteeism Level". Then removed " Chronic Absenteeism" from the filter labels. The data and the datatable does not change, just the filter checkbox labels.
filter_checkbox2("abs_levels", "Chronic Absenteeism Level",
sd, ~ abs_levels, inline = TRUE,
groupLabels = unlist(unique(ican_tab$abs_levels)) %>%
str_replace(" Chronic Absenteeism", "") %>% sort())
The first image is your table with options 1, 2, 3, and 5 (not 4).
The top version in the next image has options 1, 2, 3, and 5 (not 4). The bottom is the original table. After that
If I've left anything unclear or if have any other questions, let me know.
Columns in reactable (R package reactable) can be styled with a list of styles, a function or javascript - which works if I want to use one way only. How can these be combined (without rewriting the list, function or javascript code?
Example:
library(reactable)
list_style <- list(background = "#eee")
js_style <- JS("
function(rowInfo) {
return {fontWeight: 'bold' }
}
")
fn_style <- function(value) {
color <- "#008000"
list(color = color)
}
df <- data.frame(x = 1:10, y = 11:20)
reactable(
df,
columns = list(
x = colDef(
style = c(list_style, js_style, fn_style) # This generates the below error
)
)
)
Error:
Error in colDef(style = c(list_style, js_style, fn_style)) :
`style` must be a named list, character string, JS function, or R function
reactable(
df,
columns = list(
x = colDef(
style = list(list_style, js_style, fn_style)
)
)
)
it seems that your style out of list, please replace those code with this reactable
I am using the very interesting html widget rpivotTable. I know how to preselect variables that are added as rows or columns to the pivottable, but what I really need is a preselection of certain values of these variables.
As an example of what I mean I use the code from the vignette page:
library(rpivotTable)
data(HairEyeColor)
rpivotTable(
data = HairEyeColor, rows = "Hair",cols = "Eye", vals = "Freq",
aggregatorName = "Sum", rendererName = "Table",
sorters = "function(attr) {
var sortAs = $.pivotUtilities.sortAs;
if (attr == \"Hair\"){
return sortAs([\"Red\", \"Brown\", \"Blond\", \"Black\"]);}
}", width = "100%", height = "400px"
)
If I want, e.g. to preselect the value "Red" of the "Hair" variable, is it possible to do that in this script? Something like:
library(rpivotTable)
data(HairEyeColor)
rpivotTable(
data = HairEyeColor, rows = "Hair",cols = "Eye", vals = "Freq",
aggregatorName = "Sum", rendererName = "Table", sorters = "
function(attr) {
var sortAs = $.pivotUtilities.sortAs;
if (attr == \"Hair\") { return select([\"Red\"]); }
}", width = "100%", height = "400px"
)
I know this doesn't work but is it the way to go?
Yes, if I understand correctly, this can be accomplished with inclusions and exclusions. The format is a little funky though and requires everything to be a list.
library(rpivotTable)
data(HairEyeColor)
rpivotTable(
data = HairEyeColor,
rows = "Hair",
cols="Eye",
vals = "Freq",
aggregatorName = "Sum",
rendererName = "Table",
inclusions = list(
"Hair" = list("Red")
)
)
My data is a matrix with a variable number of columns. Also , the range of values within the matrix is variable, too.
I want to build a variable number of sliderInput, each corresponding to one column in the matrix.
The higher limit of each slider should correspond to the maxRange within the matrix.
Any suggestion how to do it in one shot?
lapply(1:ncol, function(i) {
sliderInput(
paste0('a', i),
paste0('SelectA', i),
min = min(c(1:maxRange)),
max = max(c(1:maxRange)),
value = c(1, maxRange),
step =1
)
}
)
I had broadly similar problem when I wanted to create a set of input elements corresponding to values derived from the data set (min, max list of options, etc.). Broadly speaking, I would source all the required data via global.R and then reference the concepts in the elements, so on the lines of getting min/max in a slider:
global.R
# Get dates for the slider
## Delete pointless month
dta.nom$DATE_NAME <- sub("February ", replacement = "", x = dta.nom$DATE_NAME)
## Convert to number and get min/max
dta.nom$DATE_NAME <- as.numeric(x = dta.nom$DATE_NAME)
yr.min <- min(dta.nom$DATE_NAME)
yr.max <- max(dta.nom$DATE_NAME)
Then in the slider
ui.R
# Select the dates for the data
sliderInput("sliderYears", label = h5("Years"), min = yr.min,
max = yr.max, value = c(2000, 2010), sep = "",
step = 1, animate = FALSE),
Full code is on GitHub. I'm not sure if I understood you correctly but if you are interested in dynamically connecting elements of your interface in Shiny then you can make use of the updateSelectInput. Any other problems with respect to referencing the data should be solvable with use of global code and referencing the values in your interface elements.
In case someone might bump into the same problem, here is my solution:
ui.R
....
sidebarPanel(
selectInput(
inputId = "dataName",
label = "Select your data",
choices = c("data1", "data2", "data3", "data4")
),
uiOutput(outputId = "sliders")
),
.....
server.R
.....
output$sliders <- renderUI({
numSliders <- numCols(input$dataName)
lapply(1:numSliders, function(i) {
sliderInput(
inputId = paste0('column', i),
label = paste0('Select the range for column ', i),
min = min(selectRange(input$dataName)),
max = max(selectRange(input$dataName)),
value = c(min(selectRange(input$dataName)), max(selectRange(input$dataName))),
step =1)
})
})
........
selectRange is a another function in global.R:
global.R
selectRange <- function(x){
if(x == "data1"){choices = c(1:100)}
if(x == "data2"){choices = c(1:50)}
if(x == "data3"){choices = c(1:75)}
if(x == "data4"){choices = c(1:150)}
return(choices)
}