[
https://issues.apache.org/jira/browse/IOTDB-443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17028840#comment-17028840
]
Zesong Sun commented on IOTDB-443:
----------------------------------
If you need to fix this issue in your developing branch, you can cherry pick
this commit id:
33a9604b0fde5a585ef7312687f707d1ef952ec6
> ReadOnlyMemChunk round float\double data incorrectly
> ----------------------------------------------------
>
> Key: IOTDB-443
> URL: https://issues.apache.org/jira/browse/IOTDB-443
> Project: Apache IoTDB
> Issue Type: Bug
> Components: Core/Engine
> Affects Versions: 0.9.1
> Reporter: Xiangdong Huang
> Assignee: Zesong Sun
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> In our design, only RLE and TS_2DIFF need float_precision to round the value.
> However, when we read data from memory, all float\double data will be
> rounded, even if it uses Gorilla encoding.
> e.g., suppose root.sg1.d1.s1 is double datatype and Gorilla encoding. The
> float_precision=1.
> If we insert a value (t,s1) = 1, 1.123, and select the data, we will get 1.1.
> However, if we run `flush` after insert the data, and then select the data,
> we will get 1.123.
>
> How to fix: we need to modify the init() method in ReadOnlyMemChunk to check
> the encoding method first.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
