I have a data set of this format
User
1
2
3
2
3
1
1
Now I want to add a column saying count which counts the occurrence of the user. I want output in the below format.
User Count
1 1
2 1
3 1
2 2
3 2
1 2
1 3
I have few solutions but all those solutions are somewhat slow.
My data.frame has 100,000 rows now and soon it may go up to 1 million. I need a solution which is also fast.
An option using dplyr
library(dplyr)
df1 %>%
group_by(User) %>%
mutate(Count=row_number())
# User Count
#1 1 1
#2 2 1
#3 3 1
#4 2 2
#5 3 2
#6 1 2
#7 1 3
Using sqldf
library(sqldf)
sqldf('select a.*,
count(*) as Count
from df1 a, df1 b
where a.User = b.User and b.rowid <= a.rowid
group by a.rowid')
# User Count
#1 1 1
#2 2 1
#3 3 1
#4 2 2
#5 3 2
#6 1 2
#7 1 3
This is fairly easy with ave and seq.int:
> ave(User,User, FUN= seq.int)
[1] 1 1 1 2 2 2 3
This is a common strategy and is often used when the items are adjacent to each other. The second argument is the grouping variable and in this case the first argument is really kind of a dummy argument since the only thing that it contributes is a length, and it is not a requirement for ave to have adjacent rows for the values determined within groupings.
You can use getanID from my "splitstackshape" package:
library(splitstackshape)
getanID(mydf, "User")
## User .id
## 1: 1 1
## 2: 2 1
## 3: 3 1
## 4: 2 2
## 5: 3 2
## 6: 1 2
## 7: 1 3
This is essentially an approach with "data.table" that looks something like the following:
as.data.table(mydf)[, count := seq(.N), by = "User"][]
