datafusion icon indicating copy to clipboard operation
datafusion copied to clipboard

Any plan to implement datafusion go/c++ binding?

Open xiaofan-luan opened this issue 3 years ago • 6 comments

Is your feature request related to a problem or challenge? Please describe what you are trying to do. A clear and concise description of what the problem is. Ex. I'm always frustrated when [...] (This section helps Arrow developers understand the context and why for this feature, in addition to the what)

Hi All, I'm from Milvus community(https://github.com/search?q=milvus) and we are a vector database written in Go/C++, I really love what datafusion is doing! We store our scalar data in arrow, currently we copy arrow into vector or other format for executing predication. Datafusion fits our need to filter out data from arrow, just curious if we have any plan to support datafusion in cpp~ Thanks!

Describe the solution you'd like A clear and concise description of what you want to happen.

Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

Additional context Add any other context or screenshots about the feature request here.

xiaofan-luan avatar May 08 '22 15:05 xiaofan-luan

Hi @xiaofan-luan , I have been following milvus for a long time and would love to see a synergy between the two communities :D

It shouldn't be too hard to expose our high level interfaces through C ABI for C++ and Go. For example: http://bitboom.github.io/calling-rust-from-cpp. We might need to add a couple of small high level public APIs to help pass record batches between C++ and Rust through the Arrow C Data interface.

I don't know if anyone is actively working on the C binding for datafusion right now, but if there is enough interests, we could create a new repo at https://github.com/datafusion-contrib to facilitate the collaboration.

houqp avatar May 08 '22 19:05 houqp

I don't know if anyone is actively working on the C binding for datafusion right now, but if there is enough interests, we could create a new repo at https://github.com/datafusion-contrib to facilitate the collaboration.

indeed, that C API could also benefit Java binding as well, because it'll be much better than the current JNI approach (and also in future when Java panama project is landed)

jimexist avatar May 09 '22 02:05 jimexist

@houqp great! let me know if we decide to do so, we can help on evaluating the initial version and might help on contributing it!

xiaofan-luan avatar May 09 '22 03:05 xiaofan-luan

Now we just need to wait for someone from the community to lead this effort :D

houqp avatar May 09 '22 05:05 houqp

He will do it !!

gedw99 avatar Jun 09 '23 19:06 gedw99

hello, any update on this topic? I am also interesting in how to call data fusion in go app.

codeboyyong avatar May 02 '24 18:05 codeboyyong

Also interested in a Go integration

derekperkins avatar Jun 18 '24 23:06 derekperkins

Since there is already https://github.com/datafusion-contrib/datafusion-c/tree/main, can we not directly use it in our go code? Do you have any tips or caveats to watch out for? @houqp

adarshsrivastava11 avatar Jul 10 '24 19:07 adarshsrivastava11

Now we just need to wait for someone from the community to lead this effort :D

Let's say someone is interested in looking into implementing the Go bindings, what would be the best place to start?

erezrokah avatar Jul 11 '24 13:07 erezrokah

I guess the question is actually how to use rust code in golang, and I think there is no a de facto solution yet. I found a post that maybe helpful.

jayzhan211 avatar Jul 11 '24 14:07 jayzhan211