A foray into finding out what it takes to make a database. Unabashedly different with a new take on a query language.
An immediate need exists for someone to initiate and maintain the wiki. All core information will be somewhat organized in the documentation folder.
The first part to research is the storage engine. Studies of MySQL have multiple storage engines - MyISAM, Federated and InnoDB. SQL Server has a very complex system that minimizes the number of files created.
My initial thought had been a primary file with database information. Databases might be organized by directory so that a 'Customers' database would exist under a 'Customers' directory. Several files may exist such as a database schema, table files, indexes, etc. The exact nature and organization of the file system is yet undetermined. I am looking to design a hybrid between the MySQL default file per table and the default SQL Server all in one architectures.
Requirements:
- Data generation tools to create tons of data.
- Directory structure
- File architecture
Structuring the directories and placement of files seems simple. Studying several options I am beginning with this general approach:
In the proposed directory structure, there are a minimum of 2 files:
|
.dat file structure
image to be created |
More on file structure here.