cbind 1:nrows of same ID variable value to original data.frame

Question

I have a large dataframe, where a variable id (first column) recurs with different values in the second column. My idea is to order the dataframe, to split it into a list and then lapply a function which cbinds the sequence 1:nrows(variable id) to each group. My code so far:

DF <- DF[order(DF[,1]),]
DF <- split(DF,DF[,1])
DF <- lapply(1:length(DF), function(i) cbind(DF[[i]], 1:length(DF[[i]])))

But this gives me an error: arguments imply different number of rows.

Can you elaborate?

> head(DF, n=50)
   cell     area
1     1 121.2130
2     2  81.3555
3     3  81.5862
4     4  83.6345
...
33    1 121.3270
34    2  80.7832
35    3  81.1816
36    4  83.3340

DF <- DF[order(DF$cell),]

What I want is:

> head(DF, n=50)
     cell    area counter
1       1 121.213 1
33      1 121.327 2
65      1 122.171 3
97      1 122.913 4
129     1 123.697 5
161     1 124.474 6

...and so on.

This is my code:

cell.areas.t <- function(file) {

    dat = paste(file)

    DF <- read.table(dat, col.names = c("cell","area"))
    DF <- splitstackshape::getanID(DF, "cell")[]  # thanks to akrun's answer


    ggplot2::ggplot(data = DF, aes(x = .id , y = area, color = cell)) +       
        geom_line(aes(group = cell)) + geom_point(size=0.1)
}

And the plot looks like this:

Most cells increase in area, only some decrease. This is only a first try to visualize my data, so what you can't see very well is that the areas drop down periodically due to cell division.

Additional question:

There is a problem I didn't take into account beforehand, which is that after a cell division a new cell is added to the data.frame and is handed the initial index 1 (you see in the image that all cells start from .id=1, not later), which is not what I want - it needs to inherit the index of its creation time. First thing that comes into my mind is that I could use a parsing mechanism that does this job for a newly added cell variable:

DF$.id[DF$cell != temporary.cellindex] <- max(DF$.id[DF$cell != temporary.cellindex]) 

Do you have a better idea? Thanks.

There is a boundary condition which may ease the problem: fixed number of cells at the beginning (32). Another solution would be to cut away all data before the last daughter cell is created.

Update: Additional question solved, here's the code:

cell.areas.t <- function(file) {
    dat = paste(file)
    DF <- read.table(dat, col.names = c("cell","area"))
    DF$.id <- c(0, cumsum(diff(DF$cell) < 0)) + 1L # Indexing

    title <- getwd()

    myplot <- ggplot2::ggplot(data = DF, aes(x = .id , y = area, color = factor(cell))) +
        geom_line(aes(group = cell)) + geom_line(size=0.1) + theme(legend.position="none") + ggtitle(title)

    #save the plot
    image=myplot
    ggsave(file="cell_areas_time.svg", plot=image, width=10, height=8)

}

Answer 1

We can use getanID from splitstackshape

library(splitstackshape)
getanID(DF, "cell")[]