[BEAM-367] GetFractionConsumed() inaccurate for non-uniform records

• Type: Improvement
• Status: Resolved
• Priority: P3
• Resolution: Fixed
• Fix Version/s: 0.2.0-incubating
• Component/s: io-java-gcp
• Assignee: Dan Halperin
• Reporter: Ian Zhou
• Created: 22/Jun/16 00:29
• Updated: 24/Jul/20 20:54
• Resolved: 22/Jun/16 05:30

Description

GetFractionConsumed() provides inaccurate progress updates for clustered records. For example, for a range spanning [1, 10], a cluster of records around 5 (e.g. 5.000001 ..., 5.000009) will be recorded as ~50% complete upon reading the first record, and will remain at this percentage until the final record has been read. Instead, the start of the range should be changed to the first record seen (e.g. new range [5.000001, 10]). The end of the range can be changed over time through dynamic work rebalancing.

Issue Links

• links to: GitHub Pull Request #515