Replies: 4 comments
-
|
@ives9638 yes, you are right! You concerns will be solved soon in this plan. The efficiency problem in fact also exists for LSM, and much worse. TensorBase is out of the ruins of the past. I personally think this become the new paradigm of columnar storage. Furthermore, it would be greater if you'd like to join our new effort. 🚀 😄 |
Beta Was this translation helpful? Give feedback.
-
|
We should have one RFC in some time recently. But you are interested in, we can discuss or start immediately. No need to wait any thing. |
Beta Was this translation helpful? Give feedback.
-
|
current implementation has great room to improve, but it is prepared to challenge the sky-high of OLAP. |
Beta Was this translation helpful? Give feedback.
-
|
some talked in issue #49 |
Beta Was this translation helpful? Give feedback.
-
hi @jinmingjian ，very interested
After learning the basic design and concepts of Part
original description ：
"Copa (column partition): TB is columnar, so the partition is columnar, called CoPa (short for Column Partition). The data in CoPa is only appended and unordered."
the question:
If the partition data is only appended and unordered, the read will be expanded, thereby reducing query efficiency
In addition, if the data is directly dumped, disk space is also a problem.
Beta Was this translation helpful? Give feedback.
All reactions