Hi,
Le 15/03/2021 à 00:31, Ying Zhou a écrit :
I have a question about https://github.com/apache/arrow/pull/9702
<https://github.com/apache/arrow/pull/9702> (and another future PR) over
WriterOptions and ReaderOptions that are basically code copied from the ORC project
which then got Arrowized so that the names are acceptable.
After going over Cython code for Parquet I began to wonder whether I have to
make the following changes:
1. Add ‘Orc’ to the beginning of all the enums and classes I adapted from the
ORC project e.g. arrow::adapters::orc::OrcCompressionKind instead of
arrow::adapters::orc::CompressionKind.
I'd say no.
By the way, there's already a compression enum in arrow/util/type_fwd.h,
perhaps it's enough to reuse it:
https://github.com/apache/arrow/blob/master/cpp/src/arrow/util/type_fwd.h#L42
2. Make arrow::adapters::orc::(Orc)ReaderOptions and
arrow::adapters::orc::(Orc)WriterOptions immutable and instead perform all the
mutations in their respective builder types.
Neither. Plain structs should be ok.
Take a look at CSV:
https://github.com/apache/arrow/blob/master/cpp/src/arrow/csv/options.h
Regards
Antoine.
