kou opened a new pull request, #42133: URL: https://github.com/apache/arrow/pull/42133
### Rationale for this change This is PoC. We're discussion API on the mailing list https://lists.apache.org/thread/kcpyq9npnh346pw90ljwbg0wxq6hwxxh and GH-41909. `arrow::ArrayData` not `arrow::Array` has `arrow::ArrayStatistics` to keep backward compatibility and extensibility. `arrow::Array` is immutable. So we need to set all values in constructor. The current `arrow::*Array` constructors have many arguments. Adding one more argument for `arrow::ArrayStatistics` breaks ABI and reduces usability. If `arrow::ArrayData` has `arrow::ArrayStatistics`, we can prepare `arrow::ArrayStatistics` out of constructor with `arrow::ArrayData`. Because `arrow::ArrayData` is mutable. So we can change `arrow::ArrayData` not in constructor. This supports only the following types for now: * `BooleanType` * `Int*Type` * `UInt*Type` This supports associating statistics read from Apache Parquet data to `arrow::BooleanArray`/`arrow::Int*Array`/`arrow::UInt*Array`. It's for demonstrating how to use `arrow::ArrayStatistics`. Discussions: * How do we support min/max of binary/text data? Parquet has `FLBAStatistics` (FLBA = Fixed Length Byte Array). Can we use `std::string_view` for it? If so, where do we store the original value? Can we use `std::string`? It needs a copy and copying large data will not be desired. * Should we use `arrow::Scalar` not `std::variant<bool, int8_t, ...>`? If we use `arrow::Scalar`, we can support complex value such as list and struct. But `arrow::Scalar` may be heavy. ### What changes are included in this PR? TODO ### Are these changes tested? TODO ### Are there any user-facing changes? Yes. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
