if somebody could clarify, correct or add to this: Microsoft SQL first implemented a column oriented database in 2011. So this should have fixed a lot of complaints about the inefficiencies of using SQL as financial time series database. Wouldn't this be the key consideration from going from binary files to sql? Did the other flavors of sql implement column orientation? What's the best sql book which has a depth of material on column oriented implementation? I've read about people wasting time on database implementations that turned into major rabbit holes. Many reverting to binary files for its simple efficiency ( if your a script master, I assume.)
There is a Postgres extension for column-oriented storage http://stackoverflow.com/questions/820796/open-source-column-oriented-storage-engine-for-postgresql
Column-based is not a flavor of SQL. It is a storage technique, designed to utilize less time consuming hardware functions such as (disk)reading vs writing. SQL commands are not usually affected... it is the underlying activity and performance of a given command that is different. As a generally accepted idiom in practice, row-based data excels while writing data, which is not ideal for querying write-once big-data. On the other hand, column-based tends to have blemishes when updating or inserting data. Some FREE column-based systems https://my.vertica.com/community/ http://www-01.ibm.com/software/data/db2/express-c/ http://www.monetdb.com/
I'm a noob. I'll take the MS SQL route for now. I haven't worked in Unix/Linux in over a decade and that was mostly changing file attributes. However, I did read the main man for R speak well of monetdb at stackexchange when someone asked for an enterprise grade column-based db.