>
> By the way, `df[20:50,:]` is *not* a subset of `df`, but a *copy* of a 
> subset of `df`.
> Since `df` doesn't have any column whose name is "row", how could you 
> expect the "row" to be an index.
>

Because when you get the DataFrame back, there is a column called "Row" 
that looks like an index. As a user, if it looks like an index, you will 
try to use it accordingly. And that is indeed the way it works in R and 
pandas. You have to separate the way things work internally to the way the 
user will perceive certain features.

Reply via email to