Handling of existing datasets needs to be improved

[mrdwab]

Can we:

• Use something other than the number of rows to identify whether the dataset should be re-downloaded? Perhaps the _id variable that is automatically created (and is unique).
• Add a function to pass queries to the dataset before downloading (and if yes, should we then not automatically assign the output)?
• Figure out a more acceptable way of working with the data rather than putting it in the .GlobalEnv (perhaps put it in an environment itself)?