Im having trouble finding instruction on how to preform a “warm up” that adds all relevant data to the “Hot” part of the database before making the actual querys.
Thanks for trying out MapD. Thank you for pointing this out we will make sure to add documentations for this topic.
MapD uses a smart cache mechanism which caches parts of columns(chunks) into main memory or GPU memory.
You can specify Command line argument
db-query-list and pass path to file which contains
SELECT queries that you would like to be performed. Make sure you add
USER <super-user-name> <database-name>
at the very first line of the file.
USER mapd mapd
would be the best example for the build in database and super user.
Please make sure you specify columns that you want to cache in a
where condition which does not have any other filters. That is the surest way to cache the entire column.
So a in an employee database a query to cache salary would be
select count(*) from employee where salary > 1;
This will try to fit entire salary column into hottest parts of cache. Try to keep the query as simple as possible to cache the entire columns.
Select * from employee; won’t cache anything of interest since there is nothing to “process”.