[R] create an empty data frame and then fill in it

[Duncan Murdoch]

On 09/06/2016 6:22 AM, Stefano Sofia wrote:
> Dear R list users,
> sorry for this simple question, but I already spent many efforts to solve it.
>
> I create an empty data frame called df_year like
>
> df_year <- data.frame(day=as.Date(character()), hs_MteBove=integer(), hs_MtePrata=integer(), hs_Pintura=integer(), hs_Pizzo=integer(), hs_Sassotetto=integer(), hs_Sibilla=integer(), stringsAsFactors=FALSE)
>
> and then I start to fill in it with
>
> df_year$day <- seq(as.Date("2004-11-01-00-00","%Y-%m-%d"), as.Date("2005-05-01-00-00","%Y-%m-%d"), by="day")
>
> but I get the following error:
> "replacement has 182 rows, data has 0"
>
> Where is my silly mistake?

Your dataframe has 0 rows, so you can't put a 182 row vector into the 
first column.

Unlike vectors, dataframes won't grow if you make assignments beyond the 
end of the rows.

There are at least a couple of solutions:

1.  Don't create columns until you have data ready for them.

You can wait to create the dataframe until your "day" column is ready:

df_year <- data.frame(day = seq(...))

As you compute other columns of the same length, you can add them, e.g.

df_year$hs_MteBove <- ...

2.  Create your columns with the right length from the beginning:

df_year <- data.frame(day = rep(as.Date(NA), 182), ...)

I don't like this solution as much.

Duncan Murdoch