ParquetWriter.Builder (and subclasses) instances are hard to reuse
- 主要言語
- Java
- スター
- 3.1k
- フォーク
- 1.6k
- 平均マージ
- 3日 12時間
- マージ済み PR(30日)
- 33
説明
Hi,
The builders to create ParquetWriter and AvroParquetWriter are very useful, but unfortunately their instance are very hard to reuse when you want to build multiple instance of ParquetWriter for different files but similar settings.
This is a shame because that's one of the concern solved by the builder pattern: from one builder you can build multiple things with similar parameters.
It would be useful if either:
- it was possible to create a builder from a pre-existing builder (except for the file name)
- it was possible to change the file name specified in the builder (and not have it definable only via constructor of the builder).
This would allow people to pass a set of options in the form of a builder to an object responsible of generically instantiating a PaquetWriter for multiple files for example.
**Reporter**: [Victor](https://issues.apache.org/jira/secure/ViewProfile.jspa?name=victor.noel)
**Note**: *This issue was originally created as [PARQUET-1562](https://issues.apache.org/jira/browse/PARQUET-1562). Please see the [migration documentation](https://issues.apache.org/jira/browse/PARQUET-2502) for further details.*
コントリビューションガイド
このリポジトリのコントリビューションガイドは索引されていません
調査の方向性
Start by locating ParquetWriter.Builder and the AvroParquetWriter builder subclasses. Compare the two requested reuse approaches and their builder state; done means one builder can create multiple writers for different files while preserving shared settings, with coverage for the supported builder types.
索引モデルが issue の本文から書いたものです。
評価
- 技術スタック
- java
- 領域
- data-engineering
- issue の種類
- 機能追加
- 難易度
- 4/5
- 見積もり時間
- 3〜5日
- 活発さ
- 停滞
- 明瞭さ
- おおむね明確
- 初心者へのやさしさ
- 38/100