Numbering Duplicates in R using sqldf - mysql

Question

I have a dataset which contains duplicate rows and I do wish to number them as following:
Original data set :
DF <- structure(list(pol_no = c(1L, 1L, 2L, 2L, 2L), os = c(23L, 33L,
45L, 56L, 45L), paid = c(45L, 67L, 78L, 89L, 78L)), .Names = c("pol_no",
"os", "paid"), class = "data.frame", row.names = c(NA, -5L))
which looks like this:
> DF
pol_no os paid
1 1 23 45
2 1 33 67
3 2 45 78
4 2 56 89
5 2 45 78
and I wish to number the duplicates in pol_no as follows:
pol_no os paid count
1 23 45 1
1 33 67 2
2 45 78 1
2 56 89 2
2 45 78 3
Thanks a lot in advance.
Regards,
Mansi
EDIT: added dput() output to make it reproducible and fixed formatting.

Answer 1

sqldf with RPostgreSQL
The SQL window functions of PostgreSQL faciliate solutions of this sort of problem. See FAQ#12 on the sqldf home page for more info on the use of PostgreSQL with sqldf:
library(RPostgreSQL)
library(sqldf)
sqldf('select *, rank() over (partition by "pol_no" order by CTID) count
from "DF"
order by CTID ')
sqldf with RSQLite
sqldf uses SQLite via RSQLite by default. Although SQLite lacks PostgreSQL's windowing functions the whole installation process is much simpler with SQLite since its an ordinary package install with nothing extra to do (whereas with PostgreSQL, PostgreSQL itself must be separately installed and configured). Lacking these facilities the SQL statement with SQLite is more complex although the length of the SQL statement is actually similar:
# if RPostgreSQL was previously attached & loaded then detach and & unload it
detach("package:RPostgreSQL", unload = TRUE)
sqldf("select a.*, count(*) count
from DF a, DF b
where a.pol_no = b.pol_no and b.rowid <= a.rowid group by a.rowid"
)
R's ave
Finally we show a solution that does not use sqldf at all but just core R functionality:
transform(DF, count = ave(pol_no, pol_no, FUN = seq_along))