[
https://issues.apache.org/jira/browse/CASSANDRA-5870?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15237895#comment-15237895
]
Evan Prothro commented on CASSANDRA-5870:
-----------------------------------------
It's surprising to me that this is marked as "won't fix" when time series data
is a key application for Cassandra. More alarming perhaps is to learn that COPY
doesn't support fractional seconds...
Could we possibly get more information on why this issue won't be fixed?
We've lost quite a few dev hours due to this bug (I say it is a bug since the
documentation for cqlshrc `time_format` options says it supports Python
directives, but it does not support `%f`.
> CQLSH not showing milliseconds in timestamps
> --------------------------------------------
>
> Key: CASSANDRA-5870
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5870
> Project: Cassandra
> Issue Type: Bug
> Components: Tools
> Environment: Linux
> Reporter: Ben Boule
> Priority: Minor
> Fix For: 1.2.9
>
>
> CQLSH does not include the milliseconds portion of the timestamp when
> outputting query results. For example on my system a time might be displayed
> like this:
> "2013-08-09 10:55:58-0400" for a time stored in cassandra as: 1376060158267
> We've found this extremely annoying when dealing with time series data as it
> will make records which occurred at different times appear to occur at the
> same time.
> I'm submitting a patch, the existing formatting code already has handling of
> some versions of python which do not support formatting time zones, I'm not
> sure which versions of python can format seconds+milliseconds so I attempted
> to supply something which will work with any time_format string and does not
> depend on the system library.
> The above time with the patch will format like this:
> "2013-08-09 10:55:58.267-0400"
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
