kudu-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Boris Tyukin <bo...@boristyukin.com>
Subject Re: decimals support anytime soon?
Date Thu, 14 Dec 2017 16:11:40 GMT
Hi Grant,

I am very happy to hear that, thanks for doing that! I've read the proposal
and it sounds good to me. I see there are discussions about defaults for
scale/precision but I think what you proposed to match Impala's default
makes sense to me.

We will be using Impala heavily since most of our initial use cases around
data warehousing and integration with our BI tools but we will be using
Spark as well. There are no plans to use clients directly but we might do
it to speed up bulk loads into Kudu. So for us it is important to have
decimal support for Impala / Kudu combo.

Thanks again, I feel better about it knowing that someone is actively
working on this. Please know it is a very important enhancement at least
for heatlhcare and financial data.

Boris

On Thu, Dec 14, 2017 at 10:52 AM, Grant Henke <ghenke@cloudera.com> wrote:

> Hi Boris,
>
> I am actively working on this and expect decimal support to be completed
> in the next release. I have an updated design doc here
> <https://docs.google.com/document/d/1aQ3BNO4uxVcPgmoSz7KbAjQig-Ji7ZEveGAqShiFjf0/edit?usp=sharing>
and
> will be posting patches over the next couple weeks for review. Please don't
> hesitate to give feedback to ensure the solution fits your needs.
>
> Additionally what integrations are most important to you? Are you using
> the clients directly? Impala? Spark?
>
> Thank you,
> Grant
>
> On Thu, Dec 14, 2017 at 8:25 AM, Boris Tyukin <boris@boristyukin.com>
> wrote:
>
>> https://issues.apache.org/jira/browse/KUDU-721
>>
>> Hi guys,
>>
>> I wonder if someone actively working on this Jira which has been opened
>> for a while and looks abundant. We really like to know if this is even on a
>> roadmap anytime soon as this is pretty important for our healthcare and
>> financial data applications. Float/double types just do not work for us and
>> while we have a workaround on keeping decimals in string columns it is
>> major pain in the neck due to a number of conversions one need to do every
>> time to query / process such data.
>>
>> I think this is the biggest issue on Kudu limitations list that really
>> makes me think if Kudu is the right engine for us.
>>
>> I would appreciate any comments.
>>
>> Boris
>>
>
>
>
> --
> Grant Henke
> Software Engineer | Cloudera
> grant@cloudera.com | twitter.com/gchenke | linkedin.com/in/granthenke
>

Mime
View raw message