[ https://issues.apache.org/jira/browse/IOTDB-1672?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17437833#comment-17437833 ]
Kaifeng Xue commented on IOTDB-1672: ------------------------------------ PR: https://github.com/apache/iotdb/pull/4313 > Separate AlignedInsertPlan from current InsertPlan > -------------------------------------------------- > > Key: IOTDB-1672 > URL: https://issues.apache.org/jira/browse/IOTDB-1672 > Project: Apache IoTDB > Issue Type: Improvement > Components: Core/Query, Core/Schema Manager, Core/Server > Reporter: Yukun Zhou > Assignee: Kaifeng Xue > Priority: Major > Fix For: 0.13.0 > > Attachments: image-2021-09-13-10-42-24-943.png > > Original Estimate: 32h > Remaining Estimate: 32h > > IoTDB support timeseries and aligned timeseries, in other words, unary > measurement and vector measurement. > Currently, the insert plans of these two types of timeseries is distinguished > by a "boolean isAligned" field and share the same interface and code. Thus > the related code will be confusing and difficult to understand. > For example. While MManager is processing the InsertPlan, the path and > measurement list provided by the plan represent different meaning. > # While process unary measurement, the prefix path means the entity path and > the measurement list represents different measurement of target timeseries. > # While process vector measurement, the prefix path represents the aligned > timeseries path and the measurement list represents the sub measurement of > target aligned timeseries. > !image-2021-09-13-10-42-24-943.png|width=534,height=294! > > Thus, it's better to process aligned timeseries insert plan individually and > the code will be more understandable. -- This message was sent by Atlassian Jira (v8.3.4#803005)