Looking at TestFuzzyRowFilter.java, there're several cases where mask starts with 1.
It would be much easier to diagnose if you come up with a unit test. Take a look at TestHRegionServerBulkLoad.java since you mentioned you're using bulk load in the cluster. Cheers On Thu, Mar 13, 2014 at 1:48 AM, Amit Sela <[email protected]> wrote: > On the same tables I get missing row keys for a mask in the prefix, if I > mask the second part of the key like this: > 201401\x00\x00\x00\x00\x00_product1___ > and fuzzy info: > {0,0,0,0,0,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0} > It seems to work.... > > Anyone encounter issues with masking the prefix ? Seems odd because the > Sematext example for using FuzzyRowFilter talks about masking the prefix... > > > > On Tue, Mar 11, 2014 at 11:07 AM, Amit Sela <[email protected]> wrote: > > > I can't seem to reproduce in unit test. > > The main difference is that I'm using bulk load in the cluster and Put > API > > in the unit test. > > > > > > > > On Mon, Mar 10, 2014 at 4:47 PM, Ted Yu <[email protected]> wrote: > > > >> Amit: > >> Can you put your scenario in a unit test so that it is easier to > pinpoint > >> where the issue is ? > >> > >> Thanks > >> > >> > >> On Mon, Mar 10, 2014 at 5:25 AM, Amit Sela <[email protected]> wrote: > >> > >> > My table contains keys of this kind over an entire month but the scan > >> > returns only for a some fo the days. > >> > I have 2010101-20140131 but the scan returns only for: > >> > 20140104, 20140110, 20140111, 20140118, 20140120, 20140125, 20140128 > >> > > >> > Using get or scan with no fuzzy filter works... > >> > > >> > > >> > On Mon, Mar 10, 2014 at 1:59 PM, Bharath Vissapragada < > >> > [email protected] > >> > > wrote: > >> > > >> > > Is it because you fixed "_US_product1___" part of the key? From > your > >> > > definition of filter you should get as output all keys of form > >> > > "yyyyMMdd_US_product1___". > >> > > can you share a key thats of this format and missing in the output? > >> > > > >> > > > >> > > On Mon, Mar 10, 2014 at 3:38 PM, Amit Sela <[email protected]> > >> wrote: > >> > > > >> > > > Hi all, > >> > > > I'm using HBase 0.94.12 + Hadoop 1.0.4. > >> > > > Trying to use FuzzyRowFilter looks like it's missing keys in the > >> scan. > >> > > > > >> > > > Row key structure: > >> > > > yyyyMMdd_Country_Product_Category1_Category2_ > >> > > > Where the date is mandatory and all other fields may be "". > >> > > > Examples: > >> > > > 20140101_US_product1___ > >> > > > 20140102__product1_bla__ > >> > > > 20140103_____ > >> > > > > >> > > > Supplying the filter with row key: > >> > > > \x00\x00\x00\x00\x00\x00\x00\x00_US_product1___ > >> > > > and fuzzy info: > >> > > > {1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0} > >> > > > > >> > > > Over a range of a month, although the key exists for every day in > >> the > >> > > > month, I get result only for some of the days. > >> > > > > >> > > > I tried it on another table and the same happens, I'll mention > that > >> > both > >> > > > tables have keys that start with yyyyMMdd. > >> > > > > >> > > > Anyone had a similar issue before ? I saw something in the mailing > >> list > >> > > > archives but no results there... > >> > > > > >> > > > Thanks, > >> > > > Amit. > >> > > > > >> > > > >> > > > >> > > > >> > > -- > >> > > Bharath Vissapragada > >> > > <http://www.cloudera.com> > >> > > > >> > > >> > > > > >
