Support .rx[row, col] indexing for matrices and data frames

[Wainberg]

It would be nice to support matrix_or_df.rx[1, 2] as an alternative to matrix_or_df.rx(1, 2). This would also let you assign to a matrix or data frame using matrix_or_df.rx[1, 2] = ....

As an extension, it would also be nice to support matrix_or_df.rx[:, 2] as an alternative to matrix_or_df.rx(True, 2), and matrix_or_df.rx[2, :] as an alternative to matrix_or_df.rx(2, True). Note that this would mean matrix_or_df.rx[2, :] would behave differently from matrix_or_df.rx[2], which is unexpected for Python but makes sense when you consider that matrix_or_df[2,] and matrix_or_df[2] are different in R.

[lgautier]

Yes. I once considered it but the subseting operator [ in R can accept named arguments. The R documentation for [ has"

Usage:

     x[i]
     x[i, j, ... , drop = TRUE]
     x[[i, exact = TRUE]]
     x[[i, j, ..., exact = TRUE]]
     x$name
     getElement(object, name)
     
     x[i] <- value
     x[i, j, ...] <- value
     x[[i]] <- value
     x$name <- value

Python does not allow it. Adding a named argument to a [ call is a syntax error.

I checked this again and PEP-0472 could have resolved this but was ultimately rejected (https://peps.python.org/pep-0472/). Alternative methods propose a dictionary as an additional index to have named parameters. This could be a good solution.