[
https://issues.apache.org/jira/browse/CASSANDRA-8139?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14180184#comment-14180184
]
Tyler Hobbs commented on CASSANDRA-8139:
----------------------------------------
I think we agree that having an inconsistency between what timestamps are
allowed at the protocol level and in queries is not good, but now that I think
about it, it probably makes more sense to remove the protocol-level restriction
of no negative timestamps. Even as a major version change, it doesn't make
sense to disallow something that was previously allowed unless there's a strong
benefit to doing that. Anyway, this discussion can happen in the follow-up
ticket.
Can you add a test case for this?
> The WRITETIME function returns null for negative timestamp values
> -----------------------------------------------------------------
>
> Key: CASSANDRA-8139
> URL: https://issues.apache.org/jira/browse/CASSANDRA-8139
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Reporter: Richard Bremner
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 2.0.12
>
> Attachments: 8139.txt
>
>
> Insert a column with a negative timestamp value:
> {code}
> INSERT INTO my_table (col1, col2, col3)
> VALUES ('val1', 'val2', 'val3')
> USING TIMESTAMP -1413614886750020;
> {code}
> Then attempt to read the *writetime*:
> {code}
> SELECT WRITETIME(col3) FROM my_table WHERE col1 = 'val1'
> {code}
> The result is *null*.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
