Hi,

First of all, thanks for your explanation.

I found that there are some method names in TsDigest are repeated in 
Statistics. I wonder that could I name the ByteBuffer format of Statistics also 
as the name such as digest or tsDigest in Statistics class? Do you think it is 
suitable?

Best regards,
Jack Tsai

________________________________
寄件者: Jialin Qiao <[email protected]>
寄件日期: Thursday, November 7, 2019 1:24:04 AM
收件者: [email protected] <[email protected]>
主旨: Re: Merge TsDigest into Statistics

Hi,

There is no need to add a subclass, you can just add a ByteBuffer[] field and 
related methods into Statistics.

Whenever deserializing, the type of the Statistics should already be known and 
you could construct an entity class. Therefore, abstract class is not a problem.

Best,
--
Jialin Qiao
School of Software, Tsinghua University

乔嘉林
清华大学 软件学院

> -----原始邮件-----
> 发件人: "Jack Tsai" <[email protected]>
> 发送时间: 2019-11-07 07:21:17 (星期四)
> 收件人: "[email protected]" <[email protected]>
> 抄送:
> 主题: Merge TsDigest into Statistics
>
> Hi,
>
> I am recently working on the issue 
> (https://issues.apache.org/jira/browse/IOTDB-279) which merging the TsDigest 
> into Statistics.
>
> I got a bit confusion on the description of this issue, especially for the 
> meaning of “merge”  here. I know the task is going to put the function of 
> TsDigest into Statistics. However, the Statistics class is an abstract class, 
> there are numbers of subclasses out there extend from it.
>
> I am not sure whether it is correct to make one more subclass extend from 
> Statistics to gain a ByteBuffer type for it or just add the code from 
> TsDigest to Statistics. Could the issue be claimed more specifically for me? 
> I will really appreciate that.
>
> Best regards,
> Jack Tsai
>

Reply via email to