Dear R users, For a time series, say y: y<-cumsum(rnorm(100)) # I used ur.df function (urca package) to test for unit root with/without a drift as follows: test<-ur.df(y,lags=3,type="drift") # this works for the artificial data here, but when I apply the same function to my very big data, it comes with the following error: Error in coef(summary(result))[2, 3] : subscript out of bounds Any suggestion please? Mamuash