Lars Volker has posted comments on this change. ( http://gerrit.cloudera.org:8080/11575 )
Change subject: IMPALA-6964: Track stats about column and page sizes in Parquet reader ...................................................................... Patch Set 14: (1 comment) http://gerrit.cloudera.org:8080/#/c/11575/12/be/src/exec/hdfs-scan-node-base.cc File be/src/exec/hdfs-scan-node-base.cc: http://gerrit.cloudera.org:8080/#/c/11575/12/be/src/exec/hdfs-scan-node-base.cc@962 PS12, Line 962: if (bytes_read_itr != bytes_read_per_col_.end()) { > boost:upgrade_lock doesn't really work for this use case. For upgrade_lock Thanks for the explanation. I'm out of ideas how to get this to a more RAII-style use of locks then. We could also just have a single lock around the map. This would simplify the code and remove the need for the atomic. I don't think that the lock would be very contended, given the operation is short and the granularity is reasonably large. It also helps that pages are not row-aligned. I'm curious what others think, though. If you stick with the atomic, please use our own AtomicInt from common/atomic.h. -- To view, visit http://gerrit.cloudera.org:8080/11575 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I322f9b324b6828df28e5caf79529085c43d7c817 Gerrit-Change-Number: 11575 Gerrit-PatchSet: 14 Gerrit-Owner: Sahil Takiar <stak...@cloudera.com> Gerrit-Reviewer: Impala Public Jenkins <impala-public-jenk...@cloudera.com> Gerrit-Reviewer: Lars Volker <l...@cloudera.com> Gerrit-Reviewer: Sahil Takiar <stak...@cloudera.com> Gerrit-Comment-Date: Fri, 30 Nov 2018 00:34:06 +0000 Gerrit-HasComments: Yes
