Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[VL] Remove Arrow from gluten #1434

Closed
jinchengchenghh opened this issue Apr 21, 2023 · 3 comments
Closed

[VL] Remove Arrow from gluten #1434

jinchengchenghh opened this issue Apr 21, 2023 · 3 comments
Labels
enhancement New feature or request stale stale velox backend works for Velox backend

Comments

@jinchengchenghh
Copy link
Contributor

jinchengchenghh commented Apr 21, 2023

  1. (Done)Change native row to column to velox row to column
  2. (Done)Change RangePartitioning rdd sample from ArrowWritableVector to native ColumnarToRow, compute pid change from ArrowWritableVector to native ColumnarToRow and concat pid ArrowBatch with VeloxColumnarBatch
  3. (Done)Remove CoalesceBatchesExec or implement velox CoalesceBatchesExec
  4. Arrow convert by netty to generate broadcast relation
  5. (Temp done)Remove some unused Arrow relevant class
  6. (WIP)Change velox input from ArrowStreamNode to velox new add ValueStreamNode
  7. (WIP)Refactor to add ColumnarBatchIterator to replace input arrow iterator
  8. (Optional) Benchmark replace Arrow parquet and Orc writer to Velox writer
@jinchengchenghh
Copy link
Contributor Author

jinchengchenghh commented Apr 21, 2023

Support all the velox data type in native ColumnarToRow to avoid fallback to valilla spark ColumnarToRow which will convert velox to arrow and then convert arrow to row

@jinchengchenghh
Copy link
Contributor Author

jinchengchenghh commented Apr 21, 2023

Shuffle velox to velox, remove Arrow IPC

@jinchengchenghh
Copy link
Contributor Author

Temporarily use static velox pool to convert arrow to velox, after split velox to velox, this static pool is not needed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request stale stale velox backend works for Velox backend
Projects
None yet
Development

No branches or pull requests

1 participant