[PARQUET-1820] [C++] Use a column filter hint to inform read prefetching in Arrow reads

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: cpp-4.0.0
• Component/s: parquet-cpp
• Labels: pull-request-available
• Assignee: David Li
• Reporter: David Li
• Created: 18/Mar/20 18:36
• Updated: 24/Mar/21 16:09
• Resolved: 01/May/20 18:45

Description

As a follow up to PARQUET-1698 and ARROW-7995, we should use the I/O coalescing facility (where available and enabled), in combination with a column filter hint, to compute and prefetch the exact byte ranges we will be reading (using the metadata). This should further improve performance on remote object stores like Amazon S3.

Issue Links

• Dependent: PARQUET-1698 [C++] Add reader option to pre-buffer entire serialized row group into memory
• is duplicated by: PARQUET-1698 [C++] Add reader option to pre-buffer entire serialized row group into memory
• is related to: ARROW-11601 [C++][Dataset] Expose pre-buffering in ParquetFileFormatReaderOptions
• relates to: ARROW-7995 [C++] IO: coalescing and caching read ranges
• links to: GitHub Pull Request #6744