Nice work, Barry! On Wed, May 4, 2016 at 7:05 PM, Sean Broeder <[email protected]> wrote:
> Wow! I had to read that twice :-) > > -----Original Message----- > From: Dave Birdsall [mailto:[email protected]] > Sent: Wednesday, May 4, 2016 4:44 PM > To: [email protected] > Subject: RE: [jira] [Created] (TRAFODION-1971) Redundant initialization > slows down bulk load > > Very nice find! > > -----Original Message----- > From: Liu, Ming (Ming) [mailto:[email protected]] > Sent: Wednesday, May 4, 2016 4:41 PM > To: [email protected] > Subject: 答复: [jira] [Created] (TRAFODION-1971) Redundant initialization > slows down bulk load > > Increadible!! > > -----邮件原件----- > 发件人: Barry Fritchman (JIRA) [mailto:[email protected]] > 发送时间: 2016年5月5日 2:47 > 收件人: [email protected] > 主题: [jira] [Created] (TRAFODION-1971) Redundant initialization slows down > bulk load > > Barry Fritchman created TRAFODION-1971: > ------------------------------------------ > > Summary: Redundant initialization slows down bulk load > Key: TRAFODION-1971 > URL: https://issues.apache.org/jira/browse/TRAFODION-1971 > Project: Apache Trafodion > Issue Type: Bug > Components: sql-exe > Affects Versions: 2.0-incubating > Reporter: Barry Fritchman > > > During the preparation phase of bulk load, HFIle parameter initialization > is > performed redundantly following each flush. For a 52-million row table with > fairly wide rows, a load takes 3.5 hours vs 35-40 minutes when the > unnecessary re-initialization is avoided. > > > > -- > This message was sent by Atlassian JIRA > (v6.3.4#6332) > -- Regards, --Qifan
