Read Me

------------------------------------------------------------------------------- 
                  open Optimized Storage for time series Data 
-------------------------------------------------------------------------------

1.  INTRO

	1.1 ABOUT
		OStsD is a lib for storage of temporal data in a relational database. 	
		This project is only for application developpers not for end users.

		The motivation is to store not frequently changing data (such as
		weather data, energy measurements, or some process data) but with  
		an unpredictable and potentially fast changing values (sometimes).
		So the sampling rate should be quite high but the archival must
		avoid a lot of values in order to preserve storage capacity.
		In this library it's done using two parameters associated with
		each variable : Delta V (enough value change) & Delta T (last
		archival date is enough distant).
		Data retrieval reconstitutes the temporal suite, by adding samples. 

		Data are stored in basic tables without compression. It's a choice.
		This gives the ability to access it without using my library, from
		every kind of other applications.

		This document is subject to change.

	1.2 CREDITS
		The project was created by F. Chaxel, in 2016.

2.  USAGE

	2.1 DATABASE
		Database setup must be done. Tables creation also : SQL scripts are
		given. Only the C# code creates the tables & index if required.

		The storage table OStsDVals uses a DOUBLE datatype for values. Every
		kind of basic types can be stored inside without lost.

		Alternative storage tables can be referenced for every datatype (Byte, 
		UInt16, ....) using the content of the OStsDStorage table : see the
		SQL scripts for that.
		The alternative storage tables must have the same fields than
		OStsDVals, except of course the Val datatype.  

		The user application must open the connexion, and check it's validity
		along the usage.

		A class, OSstD, is the central element for archival operations : in 
		Qt C++, C#, Java.

	2.2 CREATE VARIABLES
		In each languages a method is given for that :
			int CreateHistoricalVar (....
		but the table OStsVars can be populated by an external tool also.
		The added variable get an Id in the OStsDVars table. This primary 
		key can be given in the method call, if not the value is automatic.
		The primary key value is return.

	2.3 VALUES STORAGE
		In each languages a method is given for that :
			void InsertHistoricalValue (...
		Null values can be given.
		Only one process can stores safely a given variable, but several 
		processes can do storage and retrieval at the same time.

	2.4 SELECT
		In each languages a method is given for that :
			xxxx SelectHistoricalValues (...
		Return type is not the same in all language, but each one
		gives a clean reconstituted temporal suite. Today interpolation 
		method can be None or Linear : for instance with the database values
		(20.0 at 10:00:00) and (24.0 at 10:10:00) query requested at 10:05:00 
		gives 20.0 (Interpolation.None) or 22.0 (Interpolation.Linear).

	2.4 DELETE, UPDATE
		Do it by yourself if required.

3.  TESTS
	Tests made :
		- with SqlServer, MySql, SqlServerCe in C#
		- with SQLite in Qt
		- with Apache Derby in Java under Netbeans 	

4.  SUPPORT
	There's no support for the project at this time, and certainly never. 
	If you write to me, I'm unlikely to answer. 

5.  REPORT ERRORS
	Yes, there be errors alright. There always are. Many won't be interesting
	though. Eg. if you find a computer that behaves differently from others, 
	I'm unlikely to care. This is not a commercial project and I'm not trying 
	to push it to the greater good of the GPL world. (This may change though.)
	If you find a bug, write to me using the Sourceforge link..

6.  CONTRIBUTE
	Really? You think it's missing something? It's not really meant as a huge 
	project, but if you really must, try contacting me using the Sourceforge
	link.

7.  MISC
	Project web page is located at: 
	https://sourceforge.net/projects/ostsd/

8.  DEV. ROADMAP 
	Maybe python via ODBC. Certainly only to get back data