[jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2017-02-02 Thread Pat Ferrel (JIRA)
[ https://issues.apache.org/jira/browse/PIO-45?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pat Ferrel updated PIO-45: -- Attachment: import_handmade_simple.py import script for the data file > SelfCleaningDatasource erases all data >

[jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2017-02-02 Thread Pat Ferrel (JIRA)
[ https://issues.apache.org/jira/browse/PIO-45?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pat Ferrel updated PIO-45: -- Attachment: sample-time-window-and-downsample-data.txt simple set of data to test > SelfCleaningDatasource erases

[jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2017-02-02 Thread Pat Ferrel (JIRA)
[ https://issues.apache.org/jira/browse/PIO-45?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Pat Ferrel updated PIO-45: -- Priority: Blocker (was: Critical) > SelfCleaningDatasource erases all data >

Re: [jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2016-11-24 Thread Pat Ferrel
Cool. BTW since we have a source release, should we tag major bug fixes in the master so it will be easier to advise users how to get a fix? This would also mean pushing it to master, of course. Sort of a very lite and non-official release? We can reference commit numbers but if it doesn’t get

Re: [jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2016-11-24 Thread Alex Merritt
So it looks like earlier in the process of fixing this for JDBC I broke it for HBase. Still not quite sure why, but it appears that inserting events without eventIds is the cause of the deletion. Regardless, I just moved the event id stripping to JDBCPEvents (to fix insert errors in JDBC). Also

Re: [jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2016-11-24 Thread Pat Ferrel
True, I use HBase and saw the problem there. We’ve use the SelfCleaningDatasource in the ActionML fork successfully. It’s in the integration test for the Universal Recommender, which I was porting when I saw the test failure. I assume you are using the test integrated with the PIO release? I

Re: [jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2016-11-23 Thread Alex Merritt
I first took a quick look at the merge, and it looked like the only (minor) divergence is in JDBC. And yet, I assume you are using HBase here. As was I, when I was later able to reproduce the issue (using SelfCleaningDataSourceTest). Will aim to track down & attempt a fix today / tomorrow. Alex

Re: [jira] [Updated] (PIO-45) SelfCleaningDatasource erases all data

2016-11-21 Thread Pat Ferrel
Do you have time to look at this Alex? I may have made a mistake in merging this feature. At present any use of it erases all data. Since it is only used from templates we haven’t had one that used it except your integration test that should be merged with Apache-PIO. Can you at least run those