[ARROW-4502] [C#] Add support for zero-copy reads

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.13.0
• Component/s: C#
• Labels: performance pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/21054
• Assignee: Eric Erhardt
• Reporter: Eric Erhardt
• Created: 08/Feb/19 00:18
• Updated: 11/Jan/23 07:34
• Resolved: 07/Mar/19 15:00

Description

In the Python (and C++) API, you can create a `RecordBatchStreamReader`, and if you give it an `InputStream` that supports zero-copy reads, you can get back `RecordBatch` objects without allocating new memory and copying all the data.

There is currently no way to read Arrow RecordBatch instances without allocating new memory and copying all the data. We should enable this scenario in the C# API.

 

My proposal is to create a new `class ArrowRecordBatchReader : IArrowReader`. It's constructor will take a `ReadOnlyMemory<byte> data` parameter, and it will be able to read `RecordBatch` instances just like the existing `ArrowStreamReader`. As part of this new class, we will refactor any common code out of `ArrowStreamReader` in order for the parsing logic to be shared, where necessary.

Issue Links

• links to: GitHub Pull Request #3736

Sub-Tasks

• [C#] Update Flat Buffers code to latest version: Resolved