Assign an object in a function where the object gets created within the function

Question

I have a data frame:

m.All_Tissues <- structure(list(Sample = c("1: FL_643", "2: FL_645", "3: FL_647", "4: FL_656", "5: FL_658", "6: cKO_644", "7: cKO_646", "8: cKO_654", "9: cKO_655", "10: cKO_657", "1: FL_643", "2: FL_645", "3: FL_647", "4: FL_656", "5: FL_658", "6: cKO_644", "7: cKO_646", "8: cKO_654", "9: cKO_655", "10: cKO_657", "1: FL_643", "2: FL_645", "3: FL_647", "4: FL_656", "5: FL_658", "6: cKO_644", "7: cKO_646", "8: cKO_654", "9: cKO_655", "10: cKO_657"), Genotype = structure(c(1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L), .Label = c("miR-15/16 FL", "miR-15/16 cKO"), class = "factor"), Tissue = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("iLN", "Spleen", "Skin"), class = "factor"), variable = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = "Cells/SC/Live/CD8—,, CD4+,Freq. of Parent", class = "factor"), 
value = c(41.2, 35.5, 39.5, 33.2, 39.1, 35.5, 35.7, 33.9, 
39.7, 42.4, 23.3, 18.4, 20.9, 15.5, 19, 20.5, 22.5, 21.8, 
23.8, 24.6, 28.8, 16.9, 21.4, 19.5, 25.4, 27.1, 31.3, 28.8, 
52.8, 19)), .Names = c("Sample", "Genotype", "Tissue", "variable", "value"), row.names = 101:130, class = "data.frame")

and I would like to use a function call where I can specify objects as arguments before the objects are created to make the function more flexible.

What I have working is this:

      library(dplyr)
      library(ggplot2)          

      plot_it <- function(subsets = NULL,
                      row_add = NULL) {
    temp <- droplevels(m.All_Tissues[m.All_Tissues$Tissue %in% subsets,])
    rownames(temp) <- NULL
    df <- droplevels(temp[c(row_add),])
    rownames(df) <- NULL

    color.groups <- c("black","red")
    names(color.groups) <- unique(df$Genotype)
    shape.groups <- c(16, 1)
    names(shape.groups) <- unique(df$Genotype)

    dmax = df %>% group_by("Tissue") %>%
      summarise(value = max(value, na.rm = TRUE),
               Genotype = NA)

    ggplot(df, aes(x = Tissue, y = value, color = Genotype, shape = Genotype)) +
      geom_boxplot(position = position_dodge(width = 0.75)) +
      scale_color_manual(values = color.groups) +
      scale_shape_manual(values = shape.groups)
  }
plot_it(subsets = c("Spleen", "iLN"), row_add = c(1:20))

But I would like something that looks more like this:

plot_it <- function(data.set = NULL,
                subsets = NULL,
                group.by = NULL,
                comparison = NULL,
                row_add = NULL) {
    temp <- droplevels(data.set[data.set$group.by %in% subsets,])
    rownames(temp) <- NULL
    df <- droplevels(temp[c(row_add),])

    color.groups <- c("black","red")
    names(color.groups) <- unique(df$comparison)
    shape.groups <- c(16, 1)
    names(shape.groups) <- unique(df$comparison)

    dmax = df %>% group_by(group.by) %>%
        sumarise(value = max(value, na.rm = TRUE),
                 comparison = NA)

    ggplot(df, aes(x = group.by, y = value, color = comparison, shape = comparison)) +
      geom_boxplot(position = position_dodge(width = 0.75)) +
      scale_color_manual(values = color.groups) +
      scale_shape_manual(values = shape.groups)
}
plot_it(data.set = m.All_Tissues, subsets = c("Spleen", "iLN"), group.by = "Tissue", comparison = "Genotype", row_add = c(1:20))

I'm not sure how to go about this and would really appreciate any help or pointers in the right direction!

You might be looking for `aes_string`. You'd use that in place `aes` in your function. All variables must be strings in `aes_string`, so you'd use `y = "value"` instead of `y = value`. — aosmith, Sep 29 '17 at 23:49
But you'd need a slightly different approach for the *dplyr* part of the function. See the [programming vignette](https://cran.r-project.org/web/packages/dplyr/vignettes/programming.html) and [this answer](https://stackoverflow.com/a/44122936/2461552) — aosmith, Sep 29 '17 at 23:53

score 2 · Accepted Answer · answered Sep 29 '17 at 23:56

I renamed some arguments; removed unnecessary parts; added libraries.
But the most important part is that I call columns from data frames using dataSet[, groupBy] and in ggplot2 I use color = get(comparison) (because of that you have to specify axis and color/shape names).

plot_it <- function(dataSet, subsets, groupBy, comparison, rowAdd) {
    library(dplyr)
    library(ggplot2)

    temp <- dataSet[dataSet[, groupBy] %in% subsets, ]
    rownames(temp) <- NULL
    df <- temp[rowAdd, ]

    color.groups <- c("black","red")
    names(color.groups) <- unique(df[, comparison])
    shape.groups <- c(16, 1)
    names(shape.groups) <- unique(df[, comparison])

    # dmax <- df %>% 
    #     group_by_(groupBy) %>%
    #     summarise(value = max(value, na.rm = TRUE),
    #              comparison = NA)

    ggplot(df, aes(get(groupBy), value, 
                   color = get(comparison), shape = get(comparison))) +
      geom_boxplot(position = position_dodge(width = 0.75)) +
      labs(x = groupBy,
           color = comparison,
           shape = comparison) +
      scale_color_manual(values = color.groups) +
      scale_shape_manual(values = shape.groups)
}

plot_it(m.All_Tissues, c("Spleen", "iLN"), "Tissue", "Genotype", 1:20)

dmax part in your function does nothing

Thanks! This works very well. The dmax part is actually important for generating line segments positioned based on the maximum value of each plotted `Tissue`. I'm still having trouble getting the dmax assign to work properly by using `group_by(get(groupBy)` as it results in the column name being reassigned as "get(groupBy)" rather than "Tissue". Any suggestion for this part? — John Gagnon, Oct 01 '17 at 22:49

Assign an object in a function where the object gets created within the function

1 Answers1