diogny/csvdb

A Simple CSV database engine

CsvDb

A Csv database generator library.

The database schema generated by this library is used in my dev website diogny.com, it powers a custom CSV Database Engine.

Later I'll include the PHP classes used to read this schema.

Schema of [table].[column] index

Header

Column Index
Page Count
Flags
Key Type
Unique

Structure

Key: <127> PageSize: 16
├──Key: <64> PageSize: 16
│  ├──Key: <32> PageSize: 16
│  │  ├──PageItem, Offse: 8
│  │  └──PageItem, Offset: 272
│  └--Key: <96> PageSize: 16
│     ├──PageItem, Offset: 536
│     └──PageItem, Offset: 800
└--Key: <191> PageSize: 16
   ├──Key: <159> PageSize: 16
   │  ├──PageItem, Offset: 1056
   │  └──PageItem, Offset: 1320
   └──Key: <223> PageSize: 16
	  ├──PageItem, Offset: 1584
	  └──PageItem, Offset: 1848

And an index data with all key and its value(s).

path\bin\
	__schema.bin
	[table].data or [table].csv
	[table].pager
	[table].[index].index
	[table].[index].index.bin
	
path\
	[table].[index].index.txt
	[table].[index].index.tree.txt
	[table].[index].index.duplicates.txt
	
	[table].log

To locate a record inside the CSV data, the engine first locate the index, and then search the tree structure for a matching node entry. Then goes to the mapping page where the record is.

This way the engine can locate a csv record super fast. Currently it supports a single key entry value, and a multiple key value. So a key can have multiple values, the index support this.

No need for a database engine like MySQL, Microsoft SQL for an static relatively medium size database. Later I'll show a timechart with response times.

Later I'll add full support the formats JSON, Array of Objects, CSV.