Details
-
New Feature
-
Status: Closed
-
Major
-
Resolution: Done
-
None
Description
There are several common data models in the time series database field. One is the labeling model adopted by Influxdb, etc., which is widely used in Internet monitoring scenarios; the other is the IoT tree model used by IoTDB, which is more convenient for IoT device management.. There are certain conversion methods for tree model and label model in use, but such conversion methods are not efficient in some scenarios. Therefore, this project intends to implement a native tag model management scheme in IoTDB, and explore the combination scheme with the tree model.
The current influxdb adapter uses iotdb to store the order of tags in order to generate a unique path for the combination of tags with the same semantics during implementation. In this way, in the implementation, it is necessary to first access iotdb to obtain the order of tags, and then convert the tags into paths and generate PhysicalPlan, which results in poor performance and is more complicated to implement a tag model adapter. This work enables the influxdb adapter to support the use of native tag model metadata management modules.
Attachments
Issue Links
- links to