Thank you :-) Am I right, or using those filters have the drawback of applying them to all the rows following a necessary one instead of stopping the entire process ? If I have some small number of rows with given prefix, near the beginning of a region, will the filters make the MR task scan through all the remaining records and execute the filter methods over their data ? [If so, I should probably use it combined with a final row set to the following prefix ?, which would make the MR process iterate over at most one incorrect row .. ] (Just to be clear - I want to pack this data into TableSplit object)
cheers Piotr 2009/6/15 stack <[email protected]> > On Mon, Jun 15, 2009 at 9:45 AM, Piotr Praczyk <[email protected] > >wrote: > > > > > Does anybody maybe know, if there exists a method of finding the first > row > > larger (and smaller) in the lexycographical order than given ( not > > necessarily existing) row id ? > > > > If you open a scanner with a first row, hbase will find the first row that > matches or the next one after the provided first row. > > To stay within a range that all adheres to a particular prefix, see the > > http://hadoop.apache.org/hbase/docs/r0.19.3/api/org/apache/hadoop/hbase/filter/PrefixRowFilter.html > . > Don't forget to wrap it in a > > http://hadoop.apache.org/hbase/docs/r0.19.3/api/org/apache/hadoop/hbase/filter/WhileMatchRowFilter.html > . > See the 'description' on this page, > > http://hadoop.apache.org/hbase/docs/r0.19.3/api/org/apache/hadoop/hbase/filter/package-summary.html#package_description > , > for why WhileMatchRowFilter is needed. > > St.Ack >
