cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-9499) Introduce writeVInt method to DataOutputStreamPlus
Date Tue, 16 Jun 2015 12:15:00 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-9499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14587933#comment-14587933
] 

Sylvain Lebresne commented on CASSANDRA-9499:
---------------------------------------------

bq. with the current scheme this would mostly help values in the range of [128..250). I'm
not sure if that's worth confusing everything for.

I honestly don't see how that's confusing at all, having some {{writePositiveVint}} method
on top of just {{writeVint}}, so I do happen to think this would be worth it.

bq. we basically lose out a small amount for values in the range 64..128, and -256..-1. Everything
else we gain.

I typically see cases where that's not ideal. For instance, when serializing mutation/result
sets, we'll write stuff like the number of columns contained, the number of partions, the
number of rows in each partition, ... Those will most often be small positive numbers, and
for those, being able to pack up to 250 in a single byte would be neat.

bq. we can bias towards positive encodings, since they're more common

We can and I'm not saying it's a bad idea per se, but for what it's worth I do prefer having
us be explicit on our assumptions. More generally, we know when we use those methods what
it is we write and I think in almost all case it's trivial to know if the number is positive
or not, or even if it's very very likely small (< 100-200) or not. So I'd personally prefer
multiple (simple) encodings optimized to certain assumption (obviously, within reason) over
a single one that makes hidden assumptions that we'll forget most of the time.

But with that said, if having multiple methods sounds awful for some reason, so be it.

> Introduce writeVInt method to DataOutputStreamPlus
> --------------------------------------------------
>
>                 Key: CASSANDRA-9499
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9499
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Benedict
>            Assignee: Ariel Weisberg
>            Priority: Minor
>             Fix For: 3.0 beta 1
>
>
> CASSANDRA-8099 really could do with a writeVInt method, for both fixing CASSANDRA-9498
but also efficiently encoding timestamp/deletion deltas. It should be possible to make an
especially efficient implementation against BufferedDataOutputStreamPlus.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message