How to keep linetype spacing constant despite line size

Question

I\'ve been attempting to plot lines in either ggplot2 or grid with equal spacing between line segments when the sizes differ. However I\'ve not been succesfull so I ask you

Answer 1

This is probably not what you're looking for Teunbrand, but I guess you could convert your lines to a series of thin polygonGrobs equally spaced along the lines.

This function takes a series of x and y co-ordinates and returns a dashed line (as a single treeGrob). As per your example it returns it in normalised npc co-ordinates. You have full control over the line width, dash length and break length (though not the pattern), as well as the colour. I'm afraid the units are a bit arbitrary, and this is far from production standard, but it's fairly effective:

segmentify <- function(x, y, linewidth = 1, dash_len = 1, 
                       break_len = 1, col = "black")
{
  
  linewidth <- 0.002 * linewidth
  dash_len  <- 0.01  * dash_len
  break_len <- 0.04  * break_len

  if(length(y) != length(x)) 
    stop("x and y must be the same length")
  if(!is.numeric(x) | !is.numeric(y))
    stop("x and y must be numeric vectors")
  if(length(x) < 2)
    stop("Insufficient x, y pairs to make line.")
  
  x <- scales::rescale(x)
  y <- scales::rescale(y)
  
  n_dashes <- 0
  skip_len <- break_len + dash_len
  
   df <- list()
  for(i in seq_along(x)[-1])
  {
    x_diff          <- x[i] - x[i - 1]
    y_diff          <- y[i] - y[i - 1]
    seg_len         <- sqrt(x_diff^2 + y_diff^2)
    seg_prop        <- skip_len / seg_len
    dist_from_start <- n_dashes * skip_len
    prop_start      <- dist_from_start/seg_len
    x_start         <- x[i-1] + prop_start * x_diff
    y_len           <- y_diff * seg_prop
    x_len           <- x_diff * seg_prop
    y_start         <- y[i-1] + prop_start * y_diff
    n_breaks        <- (seg_len - dist_from_start)/skip_len
    n_dashes        <- (n_dashes + n_breaks) %% 1
    n_breaks        <- floor(n_breaks)
    
    if(n_breaks)
    {
       df[[length( df) + 1]] <- data.frame(
        x = seq(x_start, x[i], by = x_len),
        y = seq(y_start, y[i], by = y_len)
        )
       df[[length( df)]]$theta <-
        atan(rep(y_diff/x_diff, length( df[[length( df)]]$x)))
    }
  }
  
   df <- do.call(rbind,  df)
   df$x1 <-  df$x + sin( df$theta) * linewidth + cos(df$theta) * dash_len
   df$x2 <-  df$x + sin( df$theta) * linewidth - cos(df$theta) * dash_len
   df$x3 <-  df$x - sin( df$theta) * linewidth - cos(df$theta) * dash_len
   df$x4 <-  df$x - sin( df$theta) * linewidth + cos(df$theta) * dash_len
   
   df$y1 <-  df$y - cos( df$theta) * linewidth + sin(df$theta) * dash_len
   df$y2 <-  df$y - cos( df$theta) * linewidth - sin(df$theta) * dash_len
   df$y3 <-  df$y + cos( df$theta) * linewidth - sin(df$theta) * dash_len
   df$y4 <-  df$y + cos( df$theta) * linewidth + sin(df$theta) * dash_len
  
   do.call(grid::grobTree, lapply(seq(nrow(df)), function(i) {
    grid::polygonGrob(c(df$x1[i], df$x2[i], df$x3[i], df$x4[i]), 
                      c(df$y1[i], df$y2[i], df$y3[i], df$y4[i]),
              gp = gpar(col = "#00000000", lwd = 0, fill = col))
   }))

}

It's fairly straightforward to use:

set.seed(2)

x <- 1:10
y <- rnorm(10)

grid::grid.newpage()
grid::grid.draw(segmentify(x, y))

And changing the line width without affecting the spacing is just like this:

grid::grid.newpage()
grid::grid.draw(segmentify(x, y, linewidth = 3))

And you can control spacing and color like this:

grid::grid.newpage()
grid::grid.draw(segmentify(x, y, linewidth = 2, break_len = 0.5, col = "forestgreen"))