0

So basically I would like to randomly sort ~200 unique ID "numbers" (strings consisting of a letter and a number e.g. x1, x2, y40, ... ) into multiple groups of 5 without replication and with some ID "numbers" not being assigned a group.

For a minimal working example I've created 32 uniquie ID "numbers" which I would like to sort into 7 groups of 3, "using" up 21 unique ID "numbers" with 11 unique ID "numbers" being left over.

# creating 32 unique IDs
ID = c("y6","x1","x3", "x4", "x5","x12","x7", "x8", "x9","z6", "x10",   
   "y1","x11","z3","y2","y3", "y4", "y5", "y7", "y8", "y9", "y10",
   "x13","z1","x6","z2", "z4", "z5",  "z7", "z8", "z9","x2")

# and this would be how I create the first group of 3 unique ID "numbers"
sample(ID, 3, replace = FALSE, prob = NULL)
# OUTPUT [1] "x3" "x6" "y8"

So far so good, but is there a way to do this without having then to erase the chosen numbers of that group from ID and then having to select the next group of 3 and so on until I have 7 groups of 3? As this might be doable for small datasets, but not for large ones.

PS: Yes I've searched, but didn't find anything that applied and I'm a bit of an R newby and am stuck. :(

I'd really appreciate any help!

CarinaDSLR
  • 17
  • 1
  • 8
  • May be `val <- sample(ID, 21, replace=FALSE);split(val,as.numeric(gl(length(val), 3, length(val))))` – akrun Mar 16 '15 at 16:15

1 Answers1

0

I think this should do the job. Basically just randomize the vector then assign a group to it. You can split this up into a list, but I think it's better to keep it at a data.frame. 

```
ID<- sample(ID, length(ID), replace = FALSE)
ID<- data.frame(ID)
ID$groups<- ceiling(seq_along(ID$ID)/3)
```
Shorpy
  • 1,549
  • 13
  • 28