Github user xuanyuanking commented on the issue:
https://github.com/apache/spark/pull/17702
```
This approach only works if the first level glob pattern matches a lot of
directories.
```
Yep, actually in our internal usage, we leave the problem to user and they
should use first wild cast to represent most of file.
```
Maybe we should just fork the Hadoop Globber and improve it to run in
parallel.
```
Thanks for your detailed explain and guidance, I'll reconsider this and
open another PR.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
