Hi, >From the online docs:
When performing a table scan<http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/client/Scan.html> where only the row keys are needed (no families, qualifiers, values or timestamps), add a FilterList with a MUST_PASS_ALLoperator to the scanner using setFilter. The filter list should include both a FirstKeyOnlyFilter<http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/filter/FirstKeyOnlyFilter.html> and a KeyOnlyFilter<http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/filter/KeyOnlyFilter.html>. Using this filter combination will result in a worst case scenario of a RegionServer reading a single value from disk and minimal network traffic to the client for a single row. -a Sent from an iPhone On Jan 30, 2016, at 14:00, onealbao <oneal...@gmail.com<mailto:oneal...@gmail.com>> wrote: Hi, In default region split policy, it first finds largest stores, then finds largest store files, and finally get split point (midkey) of the largest store file. Is there anyway to efficiently get all row-keys of a store files? I tried to use ResultScanner with setting start/end row key, but I found the time consumption of scan (scan execution and read scan record) is at least 100 times (100 ms) slower than directly get midkey (1 ms). Actually, I just want to get all row keys in a range, and I would like to use my own policy to group some row keys together. Since all data in my table has similar size. Any suggestion is appreciated. -- View this message in context: http://apache-hbase.679495.n3.nabble.com/When-split-a-region-how-to-get-row-keys-efficiently-instead-of-using-midkey-tp4077492.html Sent from the HBase Developer mailing list archive at Nabble.com<http://nabble.com>.
