hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Samuel Guo" <guosi...@gmail.com>
Subject Is key-based operations in HBase enough?
Date Thu, 28 Aug 2008 07:52:22 GMT

In my opinion, HBase acts like a key-based system. All the data in HBase is
indexed by a row key, column key, and a timestamp. We can operate the data
in HBase by keys conveniently and fast. but sometimes, a key-based system is
not enough. for example, there is a table named users in HBase. And it
stores data like <username, age, ……>. If we want get retrieve those users
whose age is 20, I think the only way is to user a *Scanner* to go through
the whole user table to find those users. And this methode seems stupid and
slow.(If I was wrong, please let me know.)

Is there any other way to solve these kind of  value-based problems? Or, can
we lay some indecies over some columns of a table to make the value-based
operations fast? Or, can we change the underlying data organization to make
HBase act like a *BiMap* that we can retrieve data through their keys or

Hope for your reply.


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message