[ https://issues.apache.org/jira/browse/TINKERPOP-1992?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16560032#comment-16560032 ]
stephen mallette commented on TINKERPOP-1992: --------------------------------------------- Identified the problem, but not sure how to fix it - note that in {{GroupStep}} we grab the first barrier: https://github.com/apache/tinkerpop/blob/9ab5efa6772bb831238f6c9a4e46fc925a45ddaa/gremlin-core/src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/GroupStep.java#L73 and we iterate that: https://github.com/apache/tinkerpop/blob/9ab5efa6772bb831238f6c9a4e46fc925a45ddaa/gremlin-core/src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/GroupStep.java#L89 and therefore if a {{ProfileStep}} is present then we don't start its timer. as a result, when the traversal timers get normalized you end up with a negative value. > count has negative time in profile > ---------------------------------- > > Key: TINKERPOP-1992 > URL: https://issues.apache.org/jira/browse/TINKERPOP-1992 > Project: TinkerPop > Issue Type: Bug > Components: process > Affects Versions: 3.3.3, 3.2.9 > Reporter: Robert Dale > Priority: Minor > > {noformat} > gremlin> g.V().group().by().by(bothE().count()).profile() > ==>Traversal Metrics > Step Count > Traversers Time (ms) % Dur > ============================================================================================================= > TinkerGraphStep(vertex,[]) 19172 > 19172 4.957 5.24 > GroupStep(identity,[VertexStep(BOTH,edge), Prof... 1 > 1 89.696 94.76 > VertexStep(BOTH,edge) 84812 > 84812 50.723 > CountGlobalStep 19172 > 19172 -39.649 > >TOTAL - > - 94.654 - > {noformat} -- This message was sent by Atlassian JIRA (v7.6.3#76005)