You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@iotdb.apache.org by "atoildw (Jira)" <ji...@apache.org> on 2020/02/15 06:44:00 UTC

[jira] [Created] (IOTDB-481) Optimize valuefilter query

atoildw created IOTDB-481:
-----------------------------

             Summary: Optimize valuefilter query
                 Key: IOTDB-481
                 URL: https://issues.apache.org/jira/browse/IOTDB-481
             Project: Apache IoTDB
          Issue Type: Improvement
          Components: Core/Engine
            Reporter: atoildw
            Assignee: atoildw
             Fix For: 0.10.0-SNAPSHOT


The current valueFilter queries are converted to point-in-time queries, which then use different seriesReaders to read data in the same thread.

One of its optimizations is that we can fetch multiple time points at once based on batchSize, then have multiple threads query the data in the seriesReader, and finally align the data in the main thread.

We can do it in 3 steps:

1. Vectorized byTimestampReader, add getBatchValuesInTimestamps(long[] times) method

2. Vectorized TimeGenerator, modify the next() method to return long[]

3. Dataset can be optimized to query results in parallel.

 

 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)