2

Thank you to everyone who commented already! I've edited my post with better code and hopefully some clarity on what I'm trying to do. (I appreciate all of the feedback - this is my first time asking a question here!)

I have a very similar question to this one here (Random Pairings that don't Repeat) but am trying to come up with a function or piece of code that I can run in R to create the pairings. Essentially, I have a pool of employees and want to come up with a way to randomly generate pairs of employees to meet every month, with no pairs repeating in future months/running of the function. (I will need to maintain the history of previous pairings.) The catch is that each employee is assigned to a working location, and I only want matches from different location.

I've gone through a number of previous queries on randomly sampling data sets in R and am comfortable with generating a random pair from my data, or pulling out an existing working group, but it's the "generating a pair that ALWAYS comes from a different group" that's tripping me up, especially since the different groups/locations have different numbers of employees so it's hard to sort the groups evenly.

Here's my dummy data, which currently has 10 "employees". The actual data set currently has over 100 employees with more being added to the pool each month:

ID <- (1:10)
Name <- c("Sansa", "Arya", "Hodor", "Jamie", "Cersei", "Tyrion", "Jon", "Sam", "Dany", "Drogo")
Email <- c("a@a.com","b@b.com", "c@c.com", "d@d.com", "e@e.com", "f@f.com", "g@g.com", "h@h.com", "i@i.com", "j@j.com")
Location <- c("winterfell", "Winterfell", "Winterfell", "Kings Landing", "Kings Landing", "Kings Landing",
             "The Wall", "The Wall", "Essos", "Essos")


df <- data.frame(ID, Name, Email, Location)

Basically, I want to write something that would say that Sansa could be randomly paired with anyone who is not Arya or Hodor, because they're in the same location, and that Jon could be paired with anyone but Sam (i.e., anyone whose location is NOT Winterfell.) If Jon and Arya were paired up once, I would like them to not be paired up again going forward, so at that point Jon could be paired with anyone but Sam or Arya. I hope I'm making sense.

I was able to run the combn function on the ID column to generate groups of two, but it doesn't account for the historical pairings that we're trying to avoid.

Is there a way to do this in R? I've tried this one (Using R, Randomly Assigning Students Into Groups Of 4) but it wasn't quite right for my needs, again because of the historical pairings.

Thank you for any help you can provide!

wbk39
  • 21
  • 2
  • Hi, first of all, you should provide some reproducible data. There are several ways to do it: manually create a dataframe (with ```data.frame()``` for example) and copy-paste the code in your post, use some data already available in R such as ```mtcars``` or ```iris```, or take a sample of your own data and create the code to reproduce it with the ```dput``` function (and then copy-paste it here) – bretauv Nov 21 '19 at 21:12
  • Can you be more specific about what you want? e.g. provide us sample data and your expected output. Because your questions include several steps and there are many solutions for each step. You need to narrow down your question. – yusuzech Nov 21 '19 at 21:24
  • Could use a little more info before giving an answer but can you do a permutation like this `combn(1:6, 2, simplify = FALSE)` using your id column? see: https://stackoverflow.com/questions/51875414/permutation-list-of-pairs and: https://stackoverflow.com/questions/22569176/r-permutations-and-combinations-with-without-replacement-and-for-distinct-non-d – QAsena Nov 21 '19 at 21:35
  • I updated my code in the post and (hopefully) explained my question better. Thank you for the feedback!! – wbk39 Nov 22 '19 at 20:57

0 Answers0