Treasure Data's primary idea portal. 

Submit your ideas & feature requests directly to our product requirements team! We look forward to hearing from you.

Add option to tables to have an automatic unique row identifier

The 'time' column, by itself, is not sufficient to determine a unique row in Treasure Data tables.   Sometimes, even using a combination of all of the columns the entire row, is not sufficient to determine a unique row.   It would be great if there was a checkbox (or cli option) when creating the table, that automatically inserted UUID values when a new row is created.

We usually solve this problem by re-writing the new data, after the fact, into another table with a UUID column.  This makes it hard to trace back to the original "raw" data row.  It would be great to be able to have a unique identifier from the very beginning that we can carry through all of our processing.


  • Rick Otten
  • Sep 2 2016