R Generate non repeating pairs in dataframe

Question

So the purpose is to compare each ID with each other ID by taking distances.

Consider the following dataframe Df

ID AN     AW
a  white  green
b  black  yellow
c  purple gray
d  white  gray

In order to compare I need a combination looking like the following:

ID   AN     AW    ID2   AN2    AW2
a  white  green   b   black  yellow
a  white  green   c   purple gray
a  white  green   d   white  gray
b   black  yellow c   purple gray 
b   black  yellow d   white  gray
c   purple gray   d   white  gray

Basically I am trying to achieve all combinations in order to take distances between the features belonging to each ID.

Here I really do not now how to begin. Any insight? Which tools from R I could use?

using the ids: `t(combn(unique(Df$ID), 2))` – Raad Mar 16 '16 at 23:03 — Raad, Mar 16 '16 at 23:03

Raad · Accepted Answer · 2016-03-16T23:57:52.210

4

One possible solution using combn and match.

ids <- combn(unique(df$ID), 2)
data.frame(df[match(ids[1,], df$ID), ], df[match(ids[2,], df$ID), ])

#     ID     AN     AW ID.1   AN.1   AW.1
# 1    a  white  green    b  black yellow
# 1.1  a  white  green    c purple   gray
# 1.2  a  white  green    d  white   gray
# 2    b  black yellow    c purple   gray
# 2.1  b  black yellow    d  white   gray
# 3    c purple   gray    d  white   gray

edited Mar 16 '16 at 23:57

answered Mar 16 '16 at 23:26

Raad

2,675
1
13
26

Simpler: `data.frame(Df[combn(Df$ID, 2)[1,],], Df[combn(Df$ID, 2)[2,],])` – alistaire Mar 16 '16 at 23:51
Only works if the levels of the factor are in the correct order corresponding to the right row. E.g. if we add `droplevels(ids[,2])` you do not get the results you want – Raad Mar 16 '16 at 23:56
Thank you guys, now I've got some new tools in my pocket! – Saul Garcia Mar 17 '16 at 06:53
@NBATrends this works perfectly in this situation, but as I tried to implement it on a big dataframe `2328439 signatures of 11 variables`, then I get this error. `Error in combn(unique(signatures$uniqueid), 2) : n < m`. Any ideas? – Saul Garcia Mar 18 '16 at 10:55
Not sure, it might be worth opening a new question – Raad Mar 18 '16 at 11:26
It would help to see a sample of your data as well. – Raad Mar 18 '16 at 11:30

R Generate non repeating pairs in dataframe

1 Answers1

Linked