How can I sort the below data frame df to df1?
df
a1 a4 a3 a5 a2
sorted data frame
df1
a1 a2 a3 a4 a5
Asked
Active
Viewed 3,564 times
1
HaagenDaz
- 411
- 5
- 15
-
`df[ , order(names(df))]` ? – mtoto Apr 07 '16 at 09:00
-
Not sure if we can close this with [this](http://stackoverflow.com/questions/20396582/order-a-mixed-vector-numbers-with-letters) – David Arenburg Apr 07 '16 at 09:35
-
@David that's a good target – Pierre L Apr 07 '16 at 09:45
-
@PierreLafortune the dupe about sorting rows, not columns though- so not sure. – David Arenburg Apr 07 '16 at 09:49
-
Possible duplicate of [Order a "mixed" vector (numbers with letters)](https://stackoverflow.com/questions/20396582/order-a-mixed-vector-numbers-with-letters) – Ronak Shah Jul 14 '17 at 08:38
3 Answers
3
We can use mixedorder from library(gtools)
library(gtools)
df1 <- df[mixedorder(colnames(df))]
df1
# a1 a3 a9 a10
#1 1 3 1 2
#2 2 4 2 3
#3 3 5 3 4
#4 4 6 4 5
#5 5 7 5 6
data
df <- data.frame(a1 = 1:5, a10=2:6, a3 = 3:7, a9= 1:5)
akrun
- 874,273
- 37
- 540
- 662
1
In base R, assuming the numbers in the colnames don't go into double digits.
df
# a1 a4 a3 a5 a2
#1 1 4 3 5 2
df[, order(names(df))]
# a1 a2 a3 a4 a5
#1 1 2 3 4 5
mtoto
- 23,919
- 4
- 58
- 71
-
I'm not sure if this safe enough. For instance `order(c("a2", "a10"))` will give a wrong order. – David Arenburg Apr 07 '16 at 09:31
0
Assuming there is no "hole" in the numbers suffixing the columns names, you can also use dplyr:
library(dplyr)
df1 <- select(df, num_range("a", 1:4))
meriops
- 997
- 7
- 6