R 闪亮的 dplyr GROUP_BY 命令缺少条目

Question

I would like to create a dynamic data frame in shiny using the group_by command.

The number of rows of the desired table depends on rv$VAR value.

As the number of rows is different between CL =="1" and CL =="2", because some districts (010102, 010103,160101) don't have vacant housings, it does not work.

How can I display those districts with 0 in the table in order to get the same number of rows for each sort of housing ?

This is a part of my table :

PC;COUN;DISTRICT;HOUSING;CL
01:0101;  010101;     01; 1
01:0101;  010101;     02; 1
01:0101;  010101;     03; 1
01:0101;  010101;     04; 2
01:0101;  010101;     05; 1
01:0102;  010102;     01; 1
01:0102;  010102;     02; 1
01:0102;  010102;     03; 1
01:0102;  010102;     04; 1
01:0102;  010102;     05; 1
01:0103;  010103;     01; 1
01:0103;  010103;     02; 1
01:0103;  010103;     03; 1
01:0103;  010103;     04; 1
01:0103;  010103;     05; 1
15:1501;  150101;     01; 1
15:1501;  150101;     02; 2
15:1501;  150101;     03; 1
15:1501;  150101;     04; 1
15:1501;  150101;     05; 1
16:1601;  160101;     01; 1
16:1601;  160101;     02; 1
16:1601;  160101;     03; 1
16:1601;  160101;     04; 1
21:2101;  210101;     01; 1
21:2101;  210101;     02; 1
21:2101;  210101;     03; 2
21:2101;  210101;     04; 1
21:2101;  210101;     05; 2
25:2501;  250101;     01; 1
25:2501;  250101;     02; 1
25:2501;  250101;     03; 1

This is a part of the code I wrote :

selectionAcc_View  <- reactive({

if (rv$CHAMP == "DISTRICT") {

      selectionAccomodations <- reactive({
        return(filter(myTable, DISTRICT %in% rv$VAR))})

tmp <- selectionAccomodations()

dfACC <- tmp %>%
    group_by(DISTRICT) %>%
    summarize(Accomodations=n())

dfMA <- filter(tmp, CL == "1" %>%
    group_by(DISTRICT) %>%
    summarize(MA=n())

dfVH <- filter(tmp, CL == "2" %>%
    group_by(DISTRICT) %>%
    summarize(VH=n())

# Create table
df <- data.frame(

    Total_Accomodations = c(dfACC$Accomodations), # Number of Accomodations

    Main_Accomodations = c(dfMA$MA), # Number of Main Accomodations

    Vacant_Housings = c(dfVH$VH) # Number of Vacant Housings

    ) # end of data.frame

  } # end of if

df

}) # End of selectionAcc_View  <- reactive({

# Output the table
output$df <- renderDataTable(selectionAcc_View(),options = list(paging =
FALSE, ordering = FALSE,searching = FALSE,info = FALSE)) 

}) # End of shinyServer(function(input, output, session) {

Please, would you have an idea ?

Thank you very much.

Answer 1

Decided to have a look as I needed some dplyr practice anyway. But it turns out this needs to use something like tidyr (that has the functions complete and spread ) to get everything working right.

The core problem is that some of the entries end up missing due to the fact that there are no records in the original data frame for certain combinations. This is like the problem that a "FULL OUTER JOIN" addresses in SQL, as opposed to the normal left and right join behavior that leave out potential entries with no corresponding data records.

complete works with factor levels to make your output "complete" when some of the summary records do not show up due to missing data of that nature. So I had to make DISTRICT and COUN and CL into factors for this to work.

spread spread out the values in a single column into multiple columns - so converts "long" data into "wide" data.

I made a complete(ish) example out of it. Haven't rigorously tested for correctness.

library(shiny)
library(dplyr)
library(tidyr)
myTable <- read.csv(sep=";",text=
'PC;COUN;DISTRICT;HOUSING;CL
01;0101;  010101;     01; 1
01;0101;  010101;     02; 1
01;0101;  010101;     03; 1
01;0101;  010101;     04; 2
01;0101;  010101;     05; 1
01;0102;  010102;     01; 1
01;0102;  010102;     02; 1
01;0102;  010102;     03; 1
01;0102;  010102;     04; 1
01;0102;  010102;     05; 1
01;0103;  010103;     01; 1
01;0103;  010103;     02; 1
01;0103;  010103;     03; 1
01;0103;  010103;     04; 1
01;0103;  010103;     05; 1
15;1501;  150101;     01; 1
15;1501;  150101;     02; 2
15;1501;  150101;     03; 1
15;1501;  150101;     04; 1
15;1501;  150101;     05; 1
16;1601;  160101;     01; 1
16;1601;  160101;     02; 1
16;1601;  160101;     03; 1
16;1601;  160101;     04; 1
21;2101;  210101;     01; 1
21;2101;  210101;     02; 1
21;2101;  210101;     03; 2
21;2101;  210101;     04; 1
21;2101;  210101;     05; 2
25;2501;  250101;     01; 1
25;2501;  250101;     02; 1
25;2501;  250101;     03; 1')
myTable$DISTRICT <- as.factor(myTable$DISTRICT)
myTable$COUN <- as.factor(myTable$COUN)
myTable$CL <- as.factor(myTable$CL)

u <- shinyUI(fluidPage(
  titlePanel("Housing Statistics"),
  sidebarLayout(position = "left",
           sidebarPanel(h3("sidebar panel"),
                        selectInput("champmode","CHAMP Mode",c("DISTRICT","COUNTY")),
                        uiOutput("uivarselect")
                       ),
           mainPanel(h3("main panel"),
                     dataTableOutput('outdf')
                     )
             )))

s <- shinyServer(function(input,output) {

  rv <- reactiveValues(VAR = NULL,CHAMP = NULL)

  observeEvent(input$champmode,{ rv$CHAMP = input$champmode })
  observeEvent(input$varmode,{ rv$VAR = input$varmode })

  output$uivarselect <- renderUI({
    req(input$champmode)
    if (rv$CHAMP == "DISTRICT") {
        vals <- unique(as.character(myTable$DISTRICT))
     } else {
        vals <- unique(as.character(myTable$COUN))
     }
    selectInput("varmode","VAR Mode",vals)
  })


  selectionAccomodations <- reactive({
      if (rv$CHAMP == "DISTRICT") {
        return(filter(myTable,DISTRICT %in% rv$VAR))
      } else {
        return(filter(myTable,COUN %in% rv$VAR))
      }
    })

  selectionAcc_View <- reactive({
        tmp <- selectionAccomodations()
        if (nrow(tmp)==0) return(tmp) # don't process empty dataframe, just display
        tmp <- group_by(tmp,DISTRICT,COUN,CL) %>% summarize(cn = n()) %>% complete(CL)
        tmp[is.na(tmp)] <- 0 # replace NAs with zero
        df <- spread(tmp,CL,cn)
        names(df) <- c("DISTRICT","COUN","Main_Accomodations","Vacant_Housings")
        df$Total_Accomodations <- df$Main_Accomodations + df$Vacant_Housings;
        return(df)
    })

    # Output the table
  output$outdf <- renderDataTable({
      req(input$varmode) # keep from display before we are set up
      selectionAcc_View()
      },options = list(paging = F,ordering = F,searching = F,info = F))
  }
)
shinyApp(ui=u,server=s)

That yields: