The tasks of dropping columns and also of replacing one column value while brining over all others are very desirable data transformations. Codd's algebra had

The tasks of dropping columns and also of replacing one column value while brining over all others are very desirable data transformations. Codd's algebra had related features (project/extend and maybe an assign) and they showed up in Tutorial D's EXTEND keyword. It is definitely something I think is important enough to add to the rquery R SQL query generator I am working on https://winvector.github.io/rquery/ https://winvector.github.io/rquery/ ( drop_columns(), and extend_*se()'s overwriting behavior). And the R community is already familiar with such piped relation style notation in the famous dplyr package.