bvaradar commented on issue #1829:
URL: https://github.com/apache/hudi/issues/1829#issuecomment-659899321
Thanks @zuyanton for the updates. IIUC, S3 optimized committer was for
optimizing writes reducing the renames done. I might be wrong but I am
generally curious on EMR optimizations
bvaradar commented on issue #1829:
URL: https://github.com/apache/hudi/issues/1829#issuecomment-659036363
@zuyanton : This sounds like a general Spark/HMS query integration issue.
Are we seeing similar behavior when running the same query over non-hudi table
?
bvaradar commented on issue #1829:
URL: https://github.com/apache/hudi/issues/1829#issuecomment-658902047
@zuyanton : HoodieParquetInputFormat relies on hadoop-mapreduce
FileInputFormat listing implementation to perform listing. There is a knob in
base FileInputFormat to tune listing