Updating data.table by inserting new rows that are different from old rows

Question

I have two data.table(dt1 & dt2). dt1 is past product data and dt2 is present product data. I want to create a third data.table that inserts new rows from dt2 into dt1 only

Answer 1

You can stack and uniqify:

unique(rbind(dt1, dt2), by=c("Product", "Level", "Color"))

Answer 2

Another alternative is to only rbind the subset of the data which is different (avoids the creation of one big data.table which contains dt1 and dt2)

dt3 <- rbind(dt1, setDT(dt2)[!dt1, on=c("Product", "Level", "Color")])
dt3[order(Product, ReviewDate),]

Answer 3

Using merge...

d<-merge(dt1, dt2, by=c("Product","Level","Color"), all.x=T,all.y=TRUE)
d$ReviewDate <-ifelse(is.na(d$ReviewDate.x), d$ReviewDate.y, d$ReviewDate.x)
as.data.frame(select(d, 1,2,3,6))

   Product Level  Color ReviewDate
1       A     0   Blue   9/7/2016
2       A     1  Black   9/8/2016
3       B     1    Red   9/7/2016
4       C     1 Purple   9/7/2016
5       C     5  White   9/8/2016
6       D     2   Blue   9/7/2016
7       E     1  Green   9/7/2016
8       F     4 Yellow   9/7/2016
9       G     3 Orange   9/8/2016