I have a data frame in R (R Shiny) that I want to sort based on a column that contains a mix of pure numerical (1, 3, 4, ...) values as well as 'numerical-string' (2a, 2b1, 2b2, 2c, ...) values.

The data frame I have is:

I want this sorted like this:

CodePudding user response：

In case of starting with more than one digits in ColumnA, e.g., 10dxxx, 211yxxx, it is safer to separate the digits apart from the non-digit substrings (you can try sub like below), and then apply order

> with(df, df[order(as.integer(sub("\\D.*", "", ColumnA)), sub("^\\d ", "", ColumnA)), ])
  ColumnA   ColumnB
1       1   Value 1
3      2a  Value 2a
5     2b1 Value 2b1
7     2b2 Value 2b2
6      2c  Value 2c
2       3   Value 3
4       4   Value 4

Data

> dput(df)
structure(list(ColumnA = c("1", "3", "2a", "4", "2b1", "2c", 
"2b2"), ColumnB = c("Value 1", "Value 3", "Value 2a", "Value 4",
"Value 2b1", "Value 2c", "Value 2b2")), class = "data.frame", row.names = c(NA,
-7L))

CodePudding user response：

An option with mixedorder

library(dplyr)
df %>% 
  arrange(order(gtools::mixedorder(ColumnA)))
  ColumnA   ColumnB
1       1   Value 1
2      2a  Value 2a
3     2b1 Value 2b1
4     2b2 Value 2b2
5      2c  Value 2c
6       3   Value 3
7       4   Value 4