DaweiLiu created IOTDB-539:
------------------------------
Summary: reduce GroupByWithoutValueFilter unnecessary BatchData
reads
Key: IOTDB-539
URL: https://issues.apache.org/jira/browse/IOTDB-539
Project: Apache IoTDB
Issue Type: Improvement
Components: Core/Engine
Reporter: DaweiLiu
In the implementation of BatchData, use readIndex to record the current read
position, and use the reset method to set readIndex equal to 0.
But in GroupByWithoutValueFilter query, a batchData will use by multiple
aggregate method. This makes it necessary to reset each time when we use it.
If the data is:
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
and two time intervals
[1-10) [10-20]
The Batch will be cached in the interval [10-20] when we compute the [1-10).
and we should use [10, 11] data for the calculation, but the current
implementation make us to move the pointer backwards from [1, 2, 3...].
So I hope to add the getCurrentIndex, resetIndex method to BatchData to reduce
unnecessary data reads
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
