Lately, I began focusing on HBase (among the column-oriented databases). While studying the source code, one question keeps popping during my mind. Considered asking this. My real question is, exactly how a row-oriented database handles information retrieval (say a choose query) and just how different is the fact that when involves column-oriented database. And just how different these databases store data within the underlying flat files ( in the finish of day, every database uses files).

Please correct me just in case I went wrong in almost any thing about this question.

Regards, Krishna

One trouble with your question would be that the lengthy-standing database term "column oriented" continues to be appropriated (some might say "hi-jacked"!) through the NOSQL community to explain something totally different from what it really initially meant. Both meanings of "column oriented" continue to be current however they make reference to completely different DBMS items. So it's frequently helplful to explain that you simply are speaking about. Within this situation, it is the NOSQL concept of the word.

Within the original concept of a column oriented database the response to your real question is that there's no difference in the manner you retrieve information. Column store is not another data model, it's just a different sort of representation in internal storage.

Within the NOSQL community however, the word column store refers to another type of data model.

Good explanations here:

http://dbmsmusings.blogspot.com/2010/03/distinguishing-two-major-types-of_29.html