Abhishek Rawat has posted comments on this change. ( http://gerrit.cloudera.org:8080/24282 )
Change subject: IMPALA-14961: Query Profile Redaction ...................................................................... Patch Set 11: Code-Review+1 (6 comments) http://gerrit.cloudera.org:8080/#/c/24282/11/be/src/service/query-profile-redaction.cc File be/src/service/query-profile-redaction.cc: http://gerrit.cloudera.org:8080/#/c/24282/11/be/src/service/query-profile-redaction.cc@289 PS11, Line 289: const unordered_map<string_view, string_view> reverse_aliases = BuildAliasReverseMap seems unnecessary can can possibly be removed? We could simply iterate over aliases and store the key:value as value:key in alias_to_original? http://gerrit.cloudera.org:8080/#/c/24282/11/be/src/service/query-profile-redaction.cc@319 PS11, Line 319: static vector<string> CollectIdentifierContextsFromJsonProfile(const Value& node) { we're returning a vector with every recursion and doing too many allocations/deallocations. We could instead pass the vector by reference to the recursion tree. http://gerrit.cloudera.org:8080/#/c/24282/11/be/src/service/query-profile-redaction.cc@358 PS11, Line 358: static vector<string> CollectInfoStringValuesByKeys( We could pass the vector by reference down the recursion tree. http://gerrit.cloudera.org:8080/#/c/24282/11/be/src/service/query-profile-redaction.cc@395 PS11, Line 395: set<string> seen_host_tokens; We need ordering? If not we could just use a vector<string> here. We could assume that the host names are not duplicates in the profile. http://gerrit.cloudera.org:8080/#/c/24282/11/be/src/service/query-profile-redaction.cc@516 PS11, Line 516: status = BuildApplyAndTrackAliases(all_ip_tokens, "ip", redacted, alias_to_original); We're calling BuildApplyAndTrackAliases multiple times and essentially scanning the profile and copying it multiple times. We could instead build a dictionary using all the tokens and apply all the aliases in one pass doing a single copy of profile. http://gerrit.cloudera.org:8080/#/c/24282/11/be/src/service/query-profile-redaction.cc@573 PS11, Line 573: unordered_map<string, string> alias_to_original; local map seems unnecessary? You could directly pass `alias_to_original` and remove the move operation later. -- To view, visit http://gerrit.cloudera.org:8080/24282 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: If0c5b4911a64888f319f212155df6e08c1800b32 Gerrit-Change-Number: 24282 Gerrit-PatchSet: 11 Gerrit-Owner: Gokul Kolady <[email protected]> Gerrit-Reviewer: Abhishek Rawat <[email protected]> Gerrit-Reviewer: Gokul Kolady <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Jason Fehr <[email protected]> Gerrit-Reviewer: Yida Wu <[email protected]> Gerrit-Comment-Date: Sat, 16 May 2026 01:57:16 +0000 Gerrit-HasComments: Yes
