[
https://issues.apache.org/jira/browse/CASSANDRA-4450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13976705#comment-13976705
]
Sylvain Lebresne commented on CASSANDRA-4450:
---------------------------------------------
bq. Why [timestamp] is using LongType.instance instead of
TimestampType.instance?
Because TimestampType is a time in milliseconds, while [timestamp] should be in
microseconds. Besides, it's actually possible (though ill-advised unless maybe
for very very specific use casse) to use a [timestamp] that is not even a date,
so it's not like TimestampType is the absolute "true" type for that field.
> CQL3: Allow preparing the consistency level, timestamp and ttl
> --------------------------------------------------------------
>
> Key: CASSANDRA-4450
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4450
> Project: Cassandra
> Issue Type: Improvement
> Affects Versions: 1.2.0
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Labels: cql3
> Fix For: 2.0 beta 1
>
>
> It could be useful to allow the preparation of the consitency level, the
> timestamp and the ttl. I.e. to allow:
> {noformat}
> UPDATE foo SET .. USING CONSISTENCY ? AND TIMESTAMP ? AND TTL ?
> {noformat}
> A slight concern is that when preparing a statement we return the names of
> the prepared variables, but none of timestamp, ttl and consistency are
> reserved names currently, so returning those as names could conflict with a
> column name. We can either:
> * make these reserved identifier (I have to add that I'm not a fan because at
> least for "timestamp", I think that's a potentially useful and common column
> name).
> * use some specific special character to indicate those are not column names,
> like returning "[timestamp]", "[ttl]", "[consistency]".
--
This message was sent by Atlassian JIRA
(v6.2#6252)
