Increase the speed of inserting data frames to oracle tables in R using rodbc

Question

I am using RODBC to insert data frames to oracle table. I have Primary key set up on the table and if a duplicate data frame comes in it will oracle will reject it. In order to make sure I am not missing any data, first I try to insert the whole data frame, if any errors, I will try to insert each record at a time. But when I try to insert each record at a time, it is taking a very long time to finish. I was curious, has anybody done any type of work like this and what was the approach. My code is below:

tryCatch({

        ch=odbcConnect("<id>",pwd = "<password>")
        sqlSave(ch,dat, tablename="<tablename>", rownames=FALSE, append=TRUE)
        },error = function(e) 
        {
          print("unable to insert the data frame, will try by each row")   
                                    ch=odbcConnect("<id>",pwd = "<password>")

             for (k in 1:nrow(dat)) 
                {
                    j<-dat[k,]
                    tryCatch({
                            sqlSave(ch,j, tablename="<tablename>", rownames=FALSE, append=TRUE)
                    },error = function(e) {
                        print("unable to insert, duplicate values")
                    })
                }
                odbcClose(ch)
        })
odbcClose(ch)

Answer 1

Move the odbcConnect out of the per-record loop.