R merge even/odd rows

Question

I need to merge eve/odd rows on a data set.

Example:

 match team count points
 1 1 ARG 11 21
 2 1 ENG 9 19
 3 2 CAN 8 15
 4 2 USA 8 17
 5 3 FRA 7 7
 6 3 NZL 8 36

To:

 match team1 t1p t1s team2 t2p t2s
 1 1 ARG 11 21 ENG 9 19
 2 2 CAN 8 15 USA 8 17
 3 3 FRA 7 7 NZL 8 36

This is my current solution:

odd <- all_scores[seq(1, nrow(all_scores), 2),]
colnames(odd) <- c("match", "team1", "t1p", "t1s")
even <- all_scores[seq(2, nrow(all_scores), 2),]
colnames(even) <- c("match", "team2", "t2p", "t2s")
new_scores <- merge(odd, even)

Is there an easier, more space efficienct, or more R-like way to do this?

Answer 1

Have a look at the reshape function. To use the jargon, your data is in "long" format and you want to turn it to a "wide" format. Before using reshape, you will need to add a column of 1, 2, 1, 2, etc. so it can be used as the "time" column. You can do that on the fly using the transform function. In all:

new_scores <- reshape(transform(all_scores, i = 1:2),
 idvar = "match", timevar = "i", direction = "wide")
# match team.1 count.1 points.1 team.2 count.2 points.2
# 1 1 ARG 11 21 ENG 9 19
# 3 2 CAN 8 15 USA 8 17
# 5 3 FRA 7 7 NZL 8 36

• Edward Carney: The reshape is nice, but there is a way to do this without even adding another column, assuming that nrow is even and the data are in a data frame called dat1: merge(dat1[seq(1,nrow(dat1)-1,2),], dat1[seq(2,nrow(dat1),2),],by='match',all=TRUE)
• flodel: It's an interesting approach and thank you for sharing it with us. I'm just curious why you point at the fact that I add a column as if it was a bad thing? One a similar note, I could point that you wrote dat1 four times and used 8 function calls when I wrote all_scores only once and used only three function calls.
• Edward Carney: Good point w.r.t. repeated mention of dat1 and function calls. I didn't mean to imply that adding a column is a "bad thing," by any means. Mostly I did it for fun to see if it were possible to do it another way in just one line. Thanks for the reply.