Reputation: 3941
Given a data frame with index and data columns like so:
AIndex <- c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15)
AData <- c(3,5,6,7,3,2,1,2,3,4,5,6,7,8,9)
DF <- data.frame(AIndex,AData)
And given a second data frame with some overlap in the index like so:
BIndex <- c(1,4,8,11,13)
BData <- c(3,5,7,6,5)
DF2 <- data.frame(BIndex,BData)
My goal is to be able to find where the Index in A and the Index in B match up in the A data frame, and then calculate the mean of that row in A + 2 rows.
For example, the first time A and B match up in A is row# 1. So I would want to find the corresponding data point in A (3) and the next two data points (5 and 6). So the mean would be 4.6
The final result would be a new data frame that looked like this:
Index Data
1 4.6
4 4
8 3
11 6
13 8
Upvotes: 2
Views: 114
Reputation: 31181
You can do this using data.table package:
library(data.table)
setDT(DF2)[,mean(DF[BIndex:(BIndex+2),'AData']),BIndex]
# BIndex V1
#1: 1 4.666667
#2: 4 4.000000
#3: 8 3.000000
#4: 11 6.000000
#5: 13 8.000000
Upvotes: 2
Reputation: 21
I am new to R and this community. I try to follow @csgillespie's step. the result turns out to be this:
> DF2 <- data.frame(BIndex, BData)
> newInd <- merge(dat, DF2, by.x="AIndex", by.y="BIndex", all.y=T)
> newInd$newCM <- (AData[mats] + AData[mats+1] + AData[mats+2]) / 3
> newInd
AIndex AData BData newCM
1 1 3 3 4.666667
2 4 7 5 4.000000
3 8 2 7 3.000000
4 11 5 6 6.000000
5 13 7 5 8.000000
Three thanks!
Upvotes: 2
Reputation: 60492
There are a few ways of doing this. The first will step usually involves finding where the elements match:
mats = match(DF2$BIndex, DF$AIndex)
To find the means, just add up the relevant values
(AData[mats] + AData[mats+1] + AData[mats+2])/3
Upvotes: 2