How to sort a data frame by user-defined (e.g. non-alphabetic order)

13,893

Solution 1

You need to specify the levels in factor and then use order with indexing:

zz <- "chrom   start
chr2    39482
chr1    203918
chr1    198282
chrX    7839028
chr17   3874"
Data <- read.table(text=zz, header = TRUE)

library(Hmisc)
library(gdata)

Data$chrom  <- reorder.factor(Data$chrom , levels = c("chr1","chr2", "chr17", "chrX"))

Data[order(Data$chrom), ]
  chrom   start
2  chr1  203918
3  chr1  198282
1  chr2   39482
5 chr17    3874
4  chrX 7839028  

or you can use this:

> Data$chrom  <- factor(chrom , levels = c("chr1","chr2", "chr17", "chrX"))
> Data[order(Data$chrom), ]
  chrom   start
2  chr1  203918
3  chr1  198282
1  chr2   39482
5 chr17    3874
4  chrX 7839028

or use this:

> Data$chrom <- reorder(Data$chrom, new.order=c("chr1","chr2", "chr17", "chrX"))
> Data[order(Data$chrom), ]

Solution 2

Try this:

dna <- structure(list(chrom = structure(c(2L, 1L, 1L, 4L, 3L), .Label = c("chr1", 
"chr2", "chr17", "chrX"), class = c("ordered", "factor")), start = c(39482L, 
203918L, 198282L, 7839028L, 3874L)), .Names = c("chrom", "start"
), row.names = c(NA, -5L), class = "data.frame")

chrom_order <- c("chr1","chr2", "chr17", "chrX")

# Make chrom column ordered. Second term defines the order
dna$chrom <- ordered(dna$chrom, chrom_order)
dna[with(dna, order(chrom, start)),]

 chrom   start
3  chr1  198282
2  chr1  203918
1  chr2   39482
5 chr17    3874
4  chrX 7839028
Share:
13,893
biohazard
Author by

biohazard

I'm a biologist "trying" to code.

Updated on June 23, 2022

Comments

  • biohazard
    biohazard almost 2 years

    Given a data frame dna

    > dna
    chrom   start
    chr2    39482
    chr1    203918
    chr1    198282
    chrX    7839028
    chr17   3874
    

    The following code reorders dna by $chrom in alphabetical ascending order and by $start in numerical ascending order:

    > dna <- dna[with(dna, order(chrom, start)), ]
    > dna
    chrom   start
    chr1    198282
    chr1    203918
    chr17   3874
    chr2    39482
    chrX    7839028
    

    However, I would like to be able to have $chrom ordered as follows (simplified for the sake of my example here):

    chrom_order <- c("chr1","chr2", "chr17", "chrX")
    

    I am not allowed to rename stuff, for example chr1 to chr01.