Reorder data.table rows so that no two consecutive rows are the same-CodePudding

Given data with 2N rows with at least N unique rows, is there an elegant way to reorder so that each row is different from the one above it?

Here's a simple example

df <- data.table(
        a = rep(2, 4),
        b = c(2, 2, 1, 1)
    )

> df
   a b
1: 2 2
2: 2 2
3: 2 1
4: 2 1

Rows 2 breaks the rule as it is the same as row 1 which is just above it. Similarly row 4 breaks the rule.

Let's say reorder_function is the function that reorders the data, then

> reorder_function(df)
   a b
1: 2 2
2: 2 1
3: 2 2
4: 2 1

The output possibilities are not unique and it doesn't matter for my use case as long as the output satisfies the rule.

CodePudding user response：

Use rowid to "Generate unique row ids within each group", here defined by columns 'a' and 'b'. order data by the row id.

df[order(rowid(a, b))]
   a b
1: 2 2
2: 2 1
3: 2 2
4: 2 1