project_notepad.txt

Tried using LinkedList for table. about 3x slower than using hashset


Example to consider:
    Find all the living (non-deceased) Grandparents for each person in a database, given:

    Person (
        mother : Person
        father : Person
    )

    Is this done using joins? How?

    Let's start with just living maternal grandmothers -- this is one such self-join.

    To get all grandparents, do we do multiple joins to add mother and father to the resulting lists?




Comparisons with other DB systems. Main memory preferred. Alternatively, locally run.
We need to choose which systems to use to get the data for our analyses.

    MongoDB (runs locally - not main memory, persistent to disk - is or was popular for small-scale and web-app development)


Clustrix says something like:
    If you plan to run ClustrixDB in production, make sure you have 3 hosts or more to build a cluster with at least:
    48 to 64 GiB of memory
    Dual 4 Core Xeon CPUs
    RHEL v6.4 to v6.6 or CentOS v6.4 to v6.6

    They offer a good testing framework for comparison
    http://docs.clustrix.com/display/CLXDOC/Evaluating+ClustrixDB+Performance