Zhang Jiyuan

@den-crane Thanks for the suggestion. I have performed the tests as requested. ## Version Confirmation I confirmed that all nodes in the cluster are running version 25.8.12.129. ## Test Results:...

@den-crane Thanks for the confirmation. Here is the EXPLAIN output from a fast node(mt-inf-ck-dasclickhouse0116.mt, duration ~30ms, read_rows=0) in the same cluster: ``` ┌─explain───────────────────────────────────────────────────────────────────────────────────┐ 1. │ Expression ((Project names + Projection))...

> Good day, [@chuyuJY](https://github.com/chuyuJY) ! Thank you for the feedback! > > This is on our plan. Thanks for the quick response! I'm glad to hear this is on your...

I have a follow-up question: I noticed that when doing batch inserts with the same `rowbinary` format, `JDBC` seems to perform significantly better than the `client`. Could you help explain...

Thanks for the timeline and explanation! I'm using version 0.9.0. The schema includes complex types like IPv4 addresses, timezone-aware DateTime, and multiple array columns which might be contributing to the...

Thanks for the suggestion! I have already tried the POJO insert API with ClickHouse client v2, and it works great! My implementation is similar to: ```java // 1. Define POJO...

Thank you for the detailed suggestions! I've tried the parameters you mentioned (GC tuning and network buffer settings), but the improvement was minimal. However, I discovered another parameter that's much...