Counting pairs in R

Question

I have a dataset of 3 columns with Column 1 being ID(NOT UNIQUE) AND COLUMNS 2 AND 3 being a positive and negative value associated with the id respectively. I am new to R and just trying to figure out how to count the number of pairs of values associated with each id. The table and unique function is not helping since I have to count the pairs. Thanks!

Consider using `interaction`. You also need to post an example. — IRTFM, Apr 08 '15 at 20:22

Chris · Accepted Answer · 2015-04-08T23:49:32.760

1

with data.table package

library(data.table)
tdata[, list(paircount = .N) , by = c("ID","COLUMN2","COLUMN3")]

EDIT:

Based on Michael's feedback, I may have misunderstood the question.

tdata[, list(paircount = nrow((unique(.SD)))), by = "ID"]

should get you what you need.

edited Apr 08 '15 at 23:49

answered Apr 08 '15 at 20:55

Chris

6,302
1
27
54

I am pretty sure that if DataScience wanted to do this, he or she would just used `table()`... – Michael Lawrence Apr 08 '15 at 23:19
What's the difference between your method and this Arun's method http://stackoverflow.com/questions/26244685/count-every-possible-pair-of-values-in-a-column-grouped-by-multiple-columns using setorder and setkey? – skan Dec 29 '16 at 01:09

score 1 · Answer 2 · answered Apr 08 '15 at 23:06

1

I assume you want to count the number of unique pairs for each ID. As @BondedDust mentioned, use interaction:

df$pair <- with(df, interaction(COLUMN2, COLUMN3))
rowSums(xtabs(~ id + pair, df) > 0)

answered Apr 08 '15 at 23:06

Michael Lawrence

1,031
5
6

score 0 · Answer 3 · answered Apr 08 '15 at 20:26

0

Maybe try

unique(data[,c("ID", "COLUMN2", "COLUMN3")])

Or, to have results groupes by ID:

by(data = data[,c("COLUMN2", "COLUMN3")],INDICES = data$ID, FUN = unique)

answered Apr 08 '15 at 20:26

Dominic Comtois

10,230
1
39
61

Counting pairs in R

3 Answers3