I am building an app in shiny (R). At the beginning the user can upload a file to use (I am doing a sort data analysis). My goal is to be able to use files without knowing how many columns this file has, and how the data exactly looks like.
So now I have to select the columns by number, and I made a small preview app for this to select columns and then display them next to the original:
library(shiny)
ui <-fluidPage(
headerPanel("Select data"),
sidebarLayout(
sidebarPanel(
fileInput("uploadFile", "XLSX file"),
textInput('vec1', 'Choose training columns', "3,4"),
actionButton("choose","choose data")
),
mainPanel(
fluidRow(
column(6,tableOutput("data_raw")),
column(6,tableOutput("data_selected"))
)
)
)
)
server <- function(input, output) {
output$data_raw <- renderTable({
inFile <- input$uploadFile
if (is.null(inFile))
return(NULL)
data_raw <<-read.xlsx(inFile$datapath, 1)
})
observe({
if(input$choose>0){
selectvec <- as.numeric(unlist(strsplit(input$vec1,",")))
output$data_selected <- renderTable(
data_selected<- data_raw[,selectvec]
)
}
})
}
shinyApp(ui,server)
Now I would like to be able to select the columns to use on basis of their header.
It feels unnatural: changing the app while running.. but in a reactive environment.. why not?
QUESTION: How can I change the UI while it is allready running, with values originating from the input?
kind regards,
Pieter
To make me feel not as dirty for answering this...I didn't debug or handle reactives properly. But here ya go. You need to respond to the file that is uploaded on the server side, extract the column names, and append thosed to the choices in a select input that then passes down to the table function as a column filter.
upload_app <- function(){
library(shiny)
ui <- bootstrapPage(
tags$div(class = "container",
column(3,
fluidRow(
fileInput(inputId = 'user_data',
label = 'Upload Data (csv)',
multiple = FALSE,
accept = c(
'text/csv',
'text/comma-separated-values',
'text/tab-separated-values',
'text/plain',
'.csv',
'.tsv'
))
),
fluidRow(
uiOutput('column_vars')
)
),
column(9,
tableOutput('filtered_table'))
)
)
server <- function(session, input, output){
var_table <- reactive({
var_data <- input$user_data
read.csv(var_data$datapath, header = TRUE,sep = ",", quote = '')
})
output$column_vars <- renderUI({
if(!is.null(var_table())){
selectInput(inputId = 'cols',
choices = colnames(var_table()),
multiple = T,
label = "Choose Columns")
}
})
output$filtered_table <- renderTable({
if(!is.null(var_table())){
if(length(input$cols)>0){
get_these <- input$cols
new_table <- var_table()[,c(get_these)]
}else {
new_table <- var_table()
}
}else {
new_table <- data.frame(data = 'Waiting')
}
return(new_table)
})
}
shinyApp(ui, server)
}
Related
In the shiny app below I upload a dataset,so its a reactive dataset, and then I get a selectInput() with its column names. Then I should be able to select a column and change it with the textInput() after hitting the actionButton().
library(shiny)
library(shinydashboard)
library(DT)
ui <- dashboardPage(
dashboardHeader(),
dashboardSidebar(
fileInput("file1", "Choose CSV File",
accept = c(
"text/csv",
"text/comma-separated-values,text/plain",
".csv")
),
uiOutput("column"),
textInput("text", label = "Set column name", placeholder = "Enter text..."),
actionButton("sub","submit")
),
dashboardBody(
dataTableOutput("process")
)
)
server <- function(input, output) {
raw<-reactive({
inFile <- input$file1
if (is.null(inFile))
return(NULL)
read.csv(inFile$datapath, header = T)
})
output$column<-renderUI({
selectInput("col","Pick a column to change its name",
choices = colnames(raw()))
})
mydf <- reactiveValues(df = raw(), names = names(raw()))
observeEvent(input$sub, {
req(input$text)
mydf$names[mydf$names == input$col] <- input$text
names(mydf$df) <- mydf$names
updateSelectInput(inputId = "col", choices = mydf$names)
})
output$process<-renderDataTable({
mydf$df
})
}
shinyApp(ui, server)
I took your line:
mydf <- reactiveValues(df = raw(), names = names(raw()))
And replaced it with:
mydf <- reactiveValues(df = NULL, names = NULL)
observeEvent(raw(), {
if(!is.null(raw())) {
mydf$df <- raw()
mydf$names <- names(raw())
}
})
And for me it functions as how I think you wanted it. The primary issue is it would crash for me upon start. As you wrote it, it tries to access raw() immediately and put it in the reactiveValues. Thus, by making the reactiveValues NULL by default, and then only adding it in when raw() exists, it seems to work perfectly.
Here is one possible approach. Maybe there are better options. Just replace iris with your input data. It should work.
A general advice is to avoid function names as object names for reactive objects, such as raw() which is a function, but also the name of a reactive in your example. It makes things more difficult to debug and it will throw error messages which don't make sense, if you forget that raw() is also a function.
library(shiny)
library(shinydashboard)
library(DT)
ui <- dashboardPage(
dashboardHeader(),
dashboardSidebar(
# fileInput("file1", "Choose CSV File",
# accept = c(
# "text/csv",
# "text/comma-separated-values,text/plain",
# ".csv")
# ),
uiOutput("column"),
textInput("text", label = "Set column name", placeholder = "Enter text..."),
actionButton("sub","submit")
),
dashboardBody(
dataTableOutput("process")
)
)
server <- function(input, output) {
mydf <- reactiveValues(names = NULL)
raw_df <- reactive({
mydat <- iris # here goes the input file
iris_nms <- mydf$names
if(!is.null(iris_nms)) {
names(mydat) <- iris_nms
}
mydat
})
output$column <- renderUI({
selectInput("col","Pick a column to change its name",
choices = colnames(raw_df()))
})
observeEvent(input$sub, {
req(input$text)
org_names <- names(raw_df())
org_names[org_names == input$col] <- input$text
mydf$names <- org_names
})
output$process<-renderDataTable({
raw_df()
})
}
shinyApp(ui, server)
I have a simple shiny app and when I press a button a .csv file is saved in the directory where the file app.R is.
I want to be able to read this csv file and render the information in a table on my shiny app.
This is a similiar example about what I would like to do
df <- data.frame(no =c(1:3),money=c(9999:10001),penalty=c(999:1001))
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
fileInput("file1", "Choose CSV File",
accept = c(
"text/csv",
"text/comma-separated-values,text/plain",
".csv")
),
tags$hr(),
checkboxInput("header", "Header", TRUE),actionButton("sort","Do Sorting")
),
mainPanel(
tableOutput("contents"),tableOutput("sortedcontents")
)
)
)
server <- function(input, output) {
rawInputData = reactive({
rawData = input$file1
if(!is.null(rawData)) {
data = read.csv(rawData$datapath);
} else {
return(NULL);
}
});
output$contents <- renderTable({
newData = rawInputData()
if(is.null(newData))
return();
newData;
})
sorting = reactive({
if(input$sort){
newData = rawInputData()
newData$moneysort <- ifelse(newData$money >=10000, 1, 0)
newData$penaltysort <- ifelse(newData$penalty >=1000, 1, 0)
}
newData
})
output$sortedcontents <- renderTable({
newData = sorting()
if(is.null(newData))
return();
newData;
})
}
}
shinyApp(ui, server)
Instead to have the opportunity to choose the file with a fileInpunt() I would like to avoid this step and automatically check a specific directory to look for the csv called "myData.csv" and render this csv in a table.
Here is an example of just reading data from local directory and rendering in shiny.
library(shiny)
write.csv(iris, 'iris.csv')
df = read.csv('iris.csv')
shinyApp(
ui = fluidPage(
fluidRow(
column(12,
tableOutput('iris_table')
)
)
),
server = function(input, output) {
output$iris_table <- renderTable(df)
}
)
This example uses a slight modification from this shiny TableOutput reference.
I am trying to create a Shiny dashboard where the user can upload data sets, naming them, and then selecting one data set from a dropdown menu.
I dont fully understand how a user can upload a data, save it and further access it. I tried using assign(), but the dataset does not show up under ls(). My try:
library(data.table)
library(shinyWidgets)
library(shinydashboard)
if (interactive()) {
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
fileInput("file1", "Choose CSV File",
accept = c(
"text/csv",
"text/comma-separated-values,text/plain",
".csv")
),
actionButton('show_ls', 'show_ls'),
# Name data
textInput('name_data', 'Store data as' , value = ''),
# Action button to upload
actionButton('upload_data', 'Upload and save data')
),
mainPanel(
tableOutput('ls')
# textOutput('ls'))
)
)
)
server <- function(input, output, session) {
# This part uploads one row from the file, so the user can selects columns and choose name of the file
data <- reactive({
file1 <- input$file1
if(is.null(file1)){return()}
data_input <- fread(file=file1$datapath, sep=",")
updateTextInput(
session,
'name_data',
value = file1$name
)
return(data_input)
})
#### ASSIGNING THE DATA
assign_data <- eventReactive(input$upload_data, {
assign(input$name_data, data(), envir = .GlobalEnv)
})
assign_the_data <- reactive(assign_data())
output$contents <- renderTable({
if(is.null(data())) return(NULL)
data()
})
outputOptions(output, 'contents', suspendWhenHidden = FALSE)
show_ls <- eventReactive(input$show_ls, {data.frame(a = ls(.GlobalEnv))})
output$ls <- renderTable({
show_ls()
})
}
shinyApp(ui, server)
}
Using assign() can be difficult because the search path through the namespace can be difficult to predict. A better solution would be to store the datasets in a reactive values list and access them from there. You can dynamically render the dataset picker UI based on the names of the datasets stored in the reactive values list.
if (interactive()) {
ui <- fluidPage(sidebarLayout(
sidebarPanel(
fileInput(
"file1",
"Choose CSV File",
accept = c(
"text/csv",
"text/comma-separated-values,text/plain",
".csv"
)
),
# Name data
textInput('name_data', 'Store data as' , value = ''),
# Action button to upload
actionButton('upload_data', 'Upload and save data')
),
mainPanel(uiOutput("dataset_picker"),
tableOutput('selected_table')
# textOutput('ls')))
))
server <- function(input, output, session) {
# This part uploads one row from the file, so the user can selects columns and choose name of the file
dfs <- reactiveValues()
observeEvent(input$upload_data, {
req(input$file1, input$name_data)
dfs[[input$name_data]] <- read.csv(input$file1$datapath[1])
print(names(dfs))
})
output$dataset_picker <- renderUI({
req(length(dfs) > 0)
pickerInput(
"dataset_picker",
label = "Choose dataset",
choices = names(reactiveValuesToList(dfs)),
selected = NULL
)
})
output$selected_table <- renderTable({
req(input$dataset_picker)
dfs[[input$dataset_picker]]
})
}
shinyApp(ui, server)
}
I want to be able to upload multiple images with file input and display the single image selected in the UI
ui.R
library(shiny)
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
fileInput("file","Upload the file", multiple = TRUE), # fileinput() function is used to get the file upload contorl option
uiOutput("selectfile")
),
mainPanel(
uiOutput('images')
)
)
)
server.R
server <- function(input,output) {
## Side bar select input widget coming through renderUI()
# Following code displays the select input widget with the list of file loaded by the user
output$selectfile <- renderUI({
if(is.null(input$file)) {return()}
list(hr(),
helpText("Select the files for which you need to see data and summary stats"),
selectInput("Select", "Select", choices=input$file$name)
)
})
output$images <- renderImage({
if(is.null(input$file)) {return(NULL)}
for (i in 1:nrow(input$file))
{
if(input$file$name[i] == input$Select){
list(src=input$file$datapath[i],
alt= "error")
print(input$file$name[i])
print(input$file$datapath[i])
}
}
})
}
With this solution, the prints of the datapath and the name shows me the right answer but i keep getting the same error after trying to render the image: "Warning: Error in basename: a character vector argument expected".
Here is a solution using base64 encoding.
library(shiny)
library(base64enc)
ui <- fluidPage(
sidebarLayout(
sidebarPanel(
fileInput("file", "Upload the file", multiple = TRUE),
uiOutput("selectfile")
),
mainPanel(
uiOutput('image')
)
)
)
server <- function(input,output) {
output$selectfile <- renderUI({
req(input$file)
list(hr(),
helpText("Select the files for which you need to see data and summary stats"),
selectInput("Select", "Select", choices=input$file$name)
)
})
output$image <- renderUI({
req(input$Select)
i <- which(input$file$name == input$Select)
if(length(i)){
base64 <- dataURI(file = input$file$datapath[i], mime = input$file$type[i])
tags$img(src = base64, alt= "error")
}
})
}
shinyApp(ui, server)
I am very new to Shiny web app with R. I want to generate summary according to choice from checkboxGroupInput which i generated dynamically when browsing the CSV file. My problem is that when i want to convert from String to numeric that time it prints NA.
I am uploading my two files which are ui.r and server.r. I am trying since two days. If anyone help me then it will be very beneficial for me.
If i did anything wrong in my code then please suggest me right way.
ui.r
library(shiny)
library(shinythemes)
shinyUI(fluidPage(
theme = shinytheme("cyborg"),
themeSelector(),
# Application title
titlePanel("Data Analytics and Visualization Dashboard"),
sidebarLayout(
sidebarPanel(
fileInput('datafile', 'Choose CSV file',accept=c('text/csv', 'text/comma-
separated-values,text/plain')),
h5("Max file size to upload is 5 MB."),
radioButtons("sep", "Seperator", choices = c(Comma = ',', semicolon = ';',
tab = "\t", space = " " )),
#checkboxInput("header", "Header?")
br(),
h4("Select columns from CSV"),
uiOutput("toCol"),
br(),
h4("Summary"),
textOutput("sum")
# tableOutput("disp")
),
mainPanel(
numericInput("obs", "Enter the number of rows to display:", 5),
tableOutput("input_file"),
plotOutput("p")
)
)
))
server.r
library(shiny)
shinyServer(function(input, output,session) {
#This function is repsonsible for reading a csv file
output$input_file <- renderTable({
file_to_read = input$datafile
if(is.null(file_to_read))
{
return()
}
read.csv(file_to_read$datapath, sep = input$sep, nrows = input$obs))
})
#This function is repsonsible for loading in the selected file
filedata <- reactive({
infile <- input$datafile
if (is.null(infile)) {
# User has not uploaded a file yet
return(NULL)
}
read.csv(infile$datapath,nrows = input$obs)
})
#The following set of functions populate the column selectors
output$toCol <- renderUI({
df <-filedata()
if (is.null(df)) return(NULL)
items=names(df)
names(items)=items
checkboxGroupInput("to", "Columns",items)
})
observe({
# db <- gsub(",","",input$to)
# print(db)
# paste( intToUtf8(160), input$to, intToUtf8(160))
# print(summary(as.numeric(as.character( paste( " ", input$to, "
#"))))) })
print(summary(as.numeric( input$to) ))})
# output$sum <- renderPrint({
# summary(input$data.frame[,as.numeric(input$var)])
# })
# output$disp <- renderTable({
# input$to
# })
# output$summary1 <- renderPrint({
# sum <- as.numeric(as.character(input$to))
# summary(sum)
#})
})
This could be a starting point, although I dont recommend using this for a productive app, as the login-process is not really safe nor encrypted. It is based solely on text-data.
But you will have to put the ui in the server and render the page depending on the login status. So there are 2 renderUI but just 1 server-function. I dont know if you can have 2 different server-functions and redirect them. I think it all has to be in 1 server-function.
library(shiny)
username = "joe"
password = "joe123"
ui <- fluidPage(
uiOutput("ui")
)
server <- function(input, output, session) {
LOGGED <- reactiveValues(user = FALSE)
observeEvent(input$action, {
if ((input$name == username ) & (input$pass == password)) {
LOGGED$user = TRUE
} else {
LOGGED$user = FALSE
}
})
observe({
if (LOGGED$user == FALSE) {
output$ui <- renderUI({
tagList(
p(HTML("User is joe <br> and password is joe123")),
textInput("name", "Enter your username"),
passwordInput("pass", "Enter your password"),
actionButton("action", label = "Action")
)
})
} else if (LOGGED$user == TRUE) {
output$ui <- renderUI({
tagList(
h1("You are logged in.")
)
})
}
})
}
shinyApp(ui, server)
Like #Codeer said, there is no line in your code like this one summary(as.numeric(paste(input$to, “input$to”))). I edited your code, so all the uncommented lines dont appear, as its not necessary to show them.
In your example, your loading the csv file twice, which you can definitly avoid.
I moved the csv-loading into the reactive only. Then you can access the loaded file everywhere in your shiny-app. And i think in your print(summary()) statement, you're missing the data, as your only printing out the summary of the input$tovariable, which is only text and if you convert it to numeric you create NA-values.
So i rearranged your code a bit, and I think its behaving the way you intend it to.
library(shiny)
library(shinythemes)
ui <- {shinyUI(fluidPage(
theme = shinytheme("cyborg"),
themeSelector(),
titlePanel("Data Analytics and Visualization Dashboard"),
sidebarLayout(
sidebarPanel(
fileInput('datafile', 'Choose CSV file',accept=c('text/csv', 'text/comma-
separated-values,text/plain')),
h5("Max file size to upload is 5 MB."),
radioButtons("sep", "Seperator", choices = c(Comma = ',', semicolon = ';',
tab = "\t", space = " " )),
br(),
h4("Select columns from CSV"),
uiOutput("toCol"),
br(),
h4("Summary"),
textOutput("sum")
),
mainPanel(
numericInput("obs", "Enter the number of rows to display:", 5),
tableOutput("input_file"),
verbatimTextOutput("summary"),
plotOutput("p")
)
)
))}
server <- shinyServer(function(input, output,session) {
#This function is repsonsible for loading and reading a csv file
filedata <- reactive({
req(input$datafile)
infile <- input$datafile
if (is.null(infile)) {
# User has not uploaded a file yet
return(NULL)
}
read.csv(infile$datapath,nrows = input$obs, sep = input$sep)
})
output$input_file <- renderTable({
filedata()
})
#The following set of functions populate the column selectors
output$toCol <- renderUI({
df <- filedata()
if (is.null(df)) return(NULL)
items=names(df)
names(items)=items
checkboxGroupInput("to", "Columns",items)
})
output$summary <- renderPrint({
req(input$to)
data <- filedata()
print(summary(data[,input$to]))
})
})
shinyApp(ui, server)
The csv file is loaded in the reactive (filedata). In the renderTable, you just enter the reactive variable - filedata(). And in the observe, you call again the reactive variable and only print out the summary of the data in the clicked column (input$to).