Home/ Questions/Q 7515715
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-30T00:53:54+00:00

For a dataframe with three columns — $x — at http://pastebin.com/SGrRUJcA $y — at

  • 0

For a dataframe with three columns —

— I have the following code to generate a dataframe that can be used to plot a stacked bar plot:

counted <- data.frame(table(myDf$x),variable='x')
counted <- rbind(counted,data.frame(table(myDf$y),variable='y'))
counted <- rbind(counted,data.frame(table(myDf$z),variable='z'))

If I then try to sort the dataframe by its Var1 column, as so —

counted.sort <- sort_df(counted,vars="Var1")

— I get a dataframe in which column Var1 now has levels in the following order:

"1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13", "14", "15", 
"16", "17", "18", 19", "20", "21", "22", "23", "24", "25", "26", "27", "28", "29",
"30", "31", "32", "33", "34", "35", 36", 37", "39", "42", "46", "47", "53", "54", "38",
"40", "41", "43", "44", "45", "48", "49", "50"

The result distorts the x-axis of my bar plot (…,”54″, “38”, “40”, “41”, “43”,…).

How can I get counted.sort to be sorted correctly by ascending number?

NB. Plotting is done as below:

stackedBp <- ggplot(counted,aes(x=Var1,y=Freq,fill=variable))
stackedBp <- stackedBp+geom_bar(stat='identity')+scale_x_discrete('Levels')+scale_y_continuous('Frequency')
stackedBp
Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    As mentioned in a comment, R’s ?factor is confusing you.

    Let’s start with table. This command is for cross-classifying observations into different, well, categories. Implicit here is that the resulting categories are most often going to be categorical rather than numeric.

    This is why when you pass the table to data.frame the table categories are converted to a factor. Inspect the structure of counted:

    > str(counted)
'data.frame':   138 obs. of  3 variables:
 $ Var1    : Factor w/ 53 levels "1","2","3","4",..: 1 2 3 4 5 6 7 8 9 10 ...
 $ Freq    : int  126 711 1059 2079 3070 2716 2745 3329 2916 2671 ...
 $ variable: Factor w/ 3 levels "x","y","z": 1 1 1 1 1 1 1 1 1 1 ...

    There’s is an important distinction between the levels and the labels of factors. Internally, factors are always stored as integer codes, and these codes may not correspond to your labels. The internal codes are totally arbitrary, just for bookkeeping.

    The safe way to proceed is to do two coercions:

    counted$Var1 <- as.numeric(as.character(counted$Var1))

    and then explicitly convert it back to a factor in your ggplot2 code:

    stackedBp <- ggplot(counted,aes(x=factor(Var1),y=Freq,fill=variable))

    Also, you should keep in mind that sorting the data frame does nothing to change the underlying order associated with the factors in that data frame. To alter that behavior you actually need to use the ordered argument to factor when specifying the levels directly. Sorting the data frame simply changes the order in which the values are stored, not how the levels are ordered.

    • 0