r - Filtering on duplicated attributes -

i have dataframe of widgets want created predictions:

mydata<-data.frame(widget=c(1,2,3,4,5),                     type=c("a","e","c","c","d")) 

the way model works create type a:e every widget , predict. each prediction assigned attribute state model used generate prediction:

 alltypes<-merge(mydata[,-which(names(mydata)=="type")],             data.frame(type = c("a","b","c","d","e"), = true)) names(alltypes)[1]<-"widget"  alltypes$predictions<-5*1:25  attr(alltypes$predictions,"modelused")<- c(rep("model1",5),rep("model2",5),rep("model1",5),rep("model3",5),rep("model2",5    )) 

i want return predictions jsutfor original type in mydata join on original table:

library(dplyr) finaldata<- inner_join(alltypes, mydata, = c("widget", "type")) 

an examination of attributes finaldata shows there 25 attributes retured.

attr(alltypes$predictions,"modelused") 

can suggest method returning 5 attributes relevant predictions returned finaldata, or method filtering on attributes can re-add correct attributes post join?

ah works:

attr(alltypes$predictions,"modelused")[which(alltypes$type=="a")] 

or better still

originalrows<-alltypes$widget==mydata$widget &                as.character(alltypes$type)==as.character(mydata$type)  mydata<-mydata[order(mydata$type,mydata$widget),] attr(alltypes$predictions,"modelused")[which(originalrows)]