Open dangrazh opened 11 months ago
Hey! Just do a PR and we can merge
Add a few tests as well
Let me work on a proper implementation and then I'll do a PR. This might take a week or two, so bare with me.
make sure to edit the README to include the capability
Sorry, had to put this aside for a while due to other priorities.
Question on the implementation of the various new
functions in Client
:
Would you prefer to have separate new_with_batch_size
functions or to pass the batch_size as an optional parameter the the existing new
functions (as in add a function parameter batch_size: Option<usize>
to the function singatures)?
Hi
Thank for this library, this helped me a lot getting started with InfluxDB.
I have a specific use case, where I need to write thousands of data points per second to the database. In a frist test on my local machine (both the executable as well as InfluxDB are running on the same machine), inserting 2k data points into a single measurement took ~7 seconds. Reading the best practices, InfluxData states that writing batches of line protocol is more efficient than single lines. Therefore I implemented a "LineBuffer" as well as two new functions "write_point_buffered" and "write_points_buffered" into a local copy of your library. The resulting performance is ~0.235 seconds for the same set of 2k data points.
Would you be interested in discussing the topic of introducing a buffered write method into your library?