[ARROW-550] [Format] Add a TensorMessage type

• Type: New Feature
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.3.0
• Component/s: Format
• External issue URL: https://github.com/apache/arrow/issues/16187
• Assignee: Wes McKinney
• Reporter: Wes McKinney
• Created: 12/Feb/17 03:58
• Updated: 11/Jan/23 07:09
• Resolved: 24/Mar/17 15:58

Description

Since all data message types at the moment are 1-dimensional, a "tensor" message will contain an array of dimensions and an order flag (C order vs. Fortran order) to enable data to be interpreted as multiple dimensions. This is similar to multidimensional arrays in APL or Fortran or MATLAB, ndarrays in NumPy, etc.

Issue Links

• is depended upon by: ARROW-657 [Python] Write and read tensors (with zero copy) into shared memory
• is related to: ARROW-658 [C++] Implement in-memory arrow::Tensor objects