+1Hi all
A community user have observed a data loss after drop of an identity column (table empty after cmd alter table ... drop column ... on identity column, easy to reproduce)
Does any of you have already face this issue ?
Short term solution could be to use sequence instead of identity - any comment ?
Regards
Francois
+1Salut François -
In fact, I just tried it - and ran into the same issue.
People usually don't run into that because the column declared as IDENTITY is usually at least a unique key, if not the primary key, and is part of the ORDER BY clause or the segmentation criteria of the most important projections, and therefore not droppable at all. I had to create a most unusual scenario to trigger the effect.
Workaround, as you say, is an explicit sequence, and defaulting the concerned column to <sequence_name>.nextval .
Cheers -
Marco
+2Yikes! I opened a Jira for this issue: VER-59818
+2@Francois - Do you know which version of Vertica the client is using? I tested on 9.0.1-0 and this bug is there.
Already have an account? Login
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.
77 4th Avenue
Waltham, MA 02451 USA
