Improve documentation for `rows_insert()` vs `rows_append()`

[mgirlich]

It seems that people easily confuse what rows_insert() is for and use it when they actually need rows_append() (e.g. tidyverse/dbplyr#1297 or tidyverse/dbplyr#1149). It would be nice to improve the documentation (or maybe even the names) of rows_insert() and rows_append() to make it more clear when to use which function.

[wdenton]

As the creator of tidyverse/dbplyr#1297, I suggest one point that would help is a mention or example of how to add a row to a table that might have some duplicated data while not permitting a complete duplicate. If rows_insert() means everything added has to be unique, and rows_append() allows creating the same row over and over (except for the automatically created primary key, that is), then is there a dbplyr way in my example to add a green apple, without defining a primary key and without resorting to a lot of SQL? Having that in the docs would be a big help.

Thanks for noting this, and for all the excellent work on dbplyr.

[hsalberti]

Approached this issue here!
#6942