ColinLeeo commented on issue #885: URL: https://github.com/apache/tsfile/issues/885#issuecomment-5138352109
Thank you for the feedback. This issue raises a valuable point. The current implementation performs device parsing and measurement lookup every time a Tablet is written. This can become a CPU hotspot when the schema is wide or the Tablet contains relatively few rows, especially when repeatedly writing with a fixed schema. Introducing a caching mechanism is a promising optimization direction. Repeated access patterns generally exhibit strong locality in device resolution, schema lookup results, and measurement access paths. Caching these results could significantly reduce redundant computation and improve overall write throughput. We will explore this direction further and evaluate more fine-grained optimization strategies based on caching, including the appropriate cache granularity, invalidation strategy, and integration with the write path. We will also analyze representative workloads to compare the benefits and costs of different approaches. In addition, we will add benchmarks covering different write patterns to ensure that the optimization is stable and its benefits are measurable. Thank you again for raising this issue. It is very helpful in identifying and investigating this performance problem. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
