How to sort part of a data frame by multiple columns in R?

Question

Sorting has been discussed quite a bit here, but I have a particular problem. I have a data frame (df) with several columns. All the observations belong to one of three types, say C1, C2, and C3. Of the various columns, one contains a value which I will call frequency (f). Example df:

  Type  freq  Val 
   C3.   0.34.  10-A
   C1.   0.90.  4-A
   C2.   0.40.  5-B
   C1.   0.65.  3-C
   C2.   0.77.  5-D
   C3.   0.11.  5-D

I need to sort this table so that the primary key is the Type, and secondary key is the frequency. However, and here’s my problem, they need to be sorted by a particular order of type. I need them by C1, C3, C2. So the final table looks like this:

   Type  freq  Val
   C1.   0.90.  4-A
   C1.   0.65.  3-C
   C3.   0.34.  10-A
   C3.   0.11.  5-D
   C2.   0.77.  5-D
   C2.   0.40.  5-B

Is there a way to do this? To pick the particular ordering scheme of the column Type. Any suggestion is much appreciated. Thanks

akrun · Accepted Answer · 2020-04-06 17:38:54Z

1

We can convert the 'Type' to factor with levels specified in the custom order

library(dplyr)
df1 %>% 
   arrange(factor(Type, levels = c('C1.', 'C3.', 'C2.')), desc(freq))
#  Type  freq  Val
#1  C1. 0.90.  4-A
#2  C1. 0.65.  3-C
#3  C3. 0.34. 10-A
#4  C3. 0.11.  5-D
#5  C2. 0.77.  5-D
#6  C2. 0.40.  5-B

Or using data.table

library(data.table)
setDT(df1)[, Type := factor(Type,  levels = c('C1.', 'C3.', 'C2.'))]
setorder(df1, Type, -freq)

In base R, we can do

df1[order(factor(df1$Type, levels = c('C1.', 'C3.', 'C2.')), -df1$freq),]

data

df1 <- structure(list(Type = c("C3.", "C1.", "C2.", "C1.", "C2.", "C3."
), freq = c("0.34.", "0.90.", "0.40.", "0.65.", "0.77.", "0.11."
), Val = c("10-A", "4-A", "5-B", "3-C", "5-D", "5-D")), 
  class = "data.frame", row.names = c(NA, 
-6L))

edited Apr 6, 2020 at 17:38

answered Apr 6, 2020 at 17:30

akrun

891k38 gold badges590 silver badges700 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Lou_A Over a year ago

I’m working with the “dplyr” library. My final df is arranging freq correctly, but the order of the Type is yielding C1, C2, C3, and not C1, C3, C2. I wonder if there’s something missing.

akrun Over a year ago

@Orion11 If you neeed to change the column, assign the factor output to that column and then it should work

Lou_A Over a year ago

Ok. I got it to work. I was missing the "$Type" part. Thank you very much.

Collectives™ on Stack Overflow

How to sort part of a data frame by multiple columns in R?

1 Answer 1

data

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

data

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related