This depends on #12 because in order to update rows in a dataset you need to specify the key column.
Read the discussion in #15, which has a big impact on how we implement this feature.
If a key was provided to the import command, we need to look in the index for a row with that key.
if the row doesn't exist, write a new row to the feed (create row)
If the row already exists and hasn't changed, do nothing.
If the row already exists and the new values are different, write an update to the feed (update row)
Scenario:
This depends on #12 because in order to update rows in a dataset you need to specify the key column. Read the discussion in #15, which has a big impact on how we implement this feature.
If a key was provided to the import command, we need to look in the index for a row with that key.
create row
)update row
)