Github user superbobry commented on the issue:
https://github.com/apache/spark/pull/17230
Okay, one more benchmark, 4 partitions of [Consumer Complaint
Database](https://catalog.data.gov/dataset/consumer-complaint-database), more
reallistic workload than the sonnets :)
This one doesn't look as neutral:
```
Benchmark Mode Cnt Score Error Units
pipe avgt 20 1124,723 ± 267,636 ms/op
pipeFormatted avgt 20 2623,735 ± 81,431 ms/op
```
So far, I don't yet have an explanation for the differences. JMC simply
shows 80% of the time is spent decoding bytes, which is reasonable for what the
benchmark is doing, but not enough to justify the 2x slowdown.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
