load_csv functions nicely without a key, using a hash of the row data.
While nice, a useful alternative that also resolves uniquely but conveys useful information is the row number. Would suggest supporting a keyword for row number as the simplest approach to implementing this. The keyword may of course be a header and if so, dropping back to using the value in the column is sound.
load_csv
functions nicely without a key, using a hash of the row data.While nice, a useful alternative that also resolves uniquely but conveys useful information is the row number. Would suggest supporting a keyword for row number as the simplest approach to implementing this. The keyword may of course be a header and if so, dropping back to using the value in the column is sound.
A candidate key might be "row" or "row_number"