Filed https://issues.apache.org/jira/browse/HBASE-10102
________________________________ From: lars hofhansl <[email protected]> To: "[email protected]" <[email protected]>; hbase-dev <[email protected]> Sent: Friday, December 6, 2013 5:31 PM Subject: Re: HBase returns old values even with max versions = 1 + dev list Specifically: Currently the workflow in ScanQueryMatcher is something like this: 1. <versions> = min(<CF versions>, <scan version>) 2. filter by timerange 3. filter out columns (i.e. columns not specified in the scan) 4. apply customer filters 5. filter by <versions> Every KV is passed through this filtering process. What we should do is this: 1. filter by <CF versions> 2. filter by timerange 3. filter out columns (i.e. columns not specified in the scan) 4. apply customer filters 5. filter by <scan versions> The trick will be doing that efficiently. -- Lars ________________________________ From: lars hofhansl <[email protected]> To: "[email protected]" <[email protected]> Sent: Friday, December 6, 2013 5:10 PM Subject: Re: HBase returns old values even with max versions = 1 The old versions can still be around until a flush and/or compaction. During a user-level scan, HBase first filters by timerange and then counts the versions. I agree, this is counter intuitive in this case. In other cases people want to first limit by timerange, and then get x numbers of versions back. We might need to start to distinguish between the number of version configured for the column family and the number of versions configured for the scan. Mind filing a jira? Can discuss solutions there. Thanks. -- Lars ________________________________ From: Niels Basjes <[email protected]> To: user <[email protected]> Sent: Friday, December 6, 2013 8:05 AM Subject: HBase returns old values even with max versions = 1 Hi, I have the desire to find the columns that have not been updated for more than a specific time period. So I want to do a scan against the columns with a timerange. The normal behavior of HBase is that you then get the latest value in that time range (which is not what I want). As far as I understand the way HBase should work is that if you set the maximum number of versions for the values in a column family to '1' it should retain only the last value that was put into the cell. What I found is different. If I do the following commands into the hbase shell create 't1', {NAME => 'c1', VERSIONS => 1} put 't1', 'r1', 'c1', 'One', 1000 put 't1', 'r1', 'c1', 'Two', 2000 put 't1', 'r1', 'c1', 'Three', 3000 get 't1', 'r1' get 't1', 'r1' , {TIMERANGE => [0,1500]} the result is this: get 't1', 'r1' COLUMN CELL c1: timestamp=3000, value=Three 1 row(s) in 0.0780 seconds get 't1', 'r1' , {TIMERANGE => [0,1500]} COLUMN CELL c1: timestamp=1000, value=One 1 row(s) in 0.1390 seconds Why does the second query return a value even though I've set the max versions to only 1? I expect that it only 'knows' about the latest value ('Three') and thus should return an empty result in the above example. What is the correct way to obtain what I'm looking for? My current workaround is that I simply retrieve the latest value for all my columns and filter them in my application code. The HBase version I currently have installed here is HBase 0.94.6-cdh4.4.0 -- Best regards / Met vriendelijke groeten, Niels Basjes
