By the way, if you want just learning sql without complex databases as MySQL you can use library sqldf in R or sqlite. sqlite is a sql-databases which is packed in only on file. There is an addon for Firefox which you can use to connect with sqlite and perform all your sql tasks: importing csv data etc.

Since you work in Python already there are several database abstraction layer (DAL or Object-relation Mapper) libraries that you can use. The DAL hides the ugly SQL/... commands from you and you do not have to learn another language. I personally use the web framework web2py that also includes migrations for SQL databases. With migrations you can change you data models and the SQL tables will be automatically adapted in the background (with your data in them). If you need some super fast queries on large amounts of data you could also consider using a document based database like couchdb or mongodb.

All that said I have to say that databases are often overestimated. Use the Python pickle module or even better JSON files as long as you can.

Read the source code! the text only if you really need to know in detail why things work

http://www.web2py.com/init/default/examples#database_examples Example: 29

http://web2py.com/books/default/chapter/29/06#select

http://www.rmunn.com/sqlalchemy-tutorial/tutorial.html quite similar to web2pys

http://pythonhosted.org/CouchDB/getting-started.html

just some hints, but I hope you get the idea

Thank you for the information. I will take a look at your suggestions and start coding!

I would suggest to use the MariaDB-MySQL instead of Oracle-MySQL. As non-professional programmer I found MariaDB easier to install than Oracle-MySQL. MariaDB is a clone (one of many) of MySQL and is fostered by one of the founders of MySQL who left Oracle-MySQL whatever the reason was. By the way, Wikipedia changed from MySQL to MariaDB. Anyway, once installed you won't find much difference and for your and my purposes there is no technical reason why to take the one or the other. For managing and sending explorative adhoc sql queries I use HeidiSQL as GUI. You can have it on Windows and Linux (with wine). If you never programmed sql I would suggest to start with MariaDB Crash Course by Ben Forta. I don't have it but I looked at the content and it seems very well explained. SQL is a wonderful language and easy to learn. You will encounter problems when you import data (due to, for exampel, control characters in flat files). When you get rid of these problems there will be only minor problems using MySQL/MariaDB. I use ODBC to connect MariaDB/MySQL with R (ODBC-Driver you can have it from the Oracle-MySQL Homepage). The MySQL Reference in the Oracle Homepage and all MySQL Books are also valid for MariaDB. If you want to learn some sql directly in R than you can use library(sqldf) which is an embedded sqlite databases in R with similar syntax. The sql you can apply on data frames. I often use sql instead of apply, lapply, etc. for some aggregation since I find sql much easier to use. Some words on object orientation: it is possible to build a framework with Django (Python), Catalyst (Perl) or what ever. For my purposes: just administration of datasets and some processing and then doing further processing and analysis in R the object oriented way would be to take a sledgehammer to crack a nut. But maybe I'm wrong. When I had to analyse data from several hundered of Excel workbooks I used Perl to cope with this issues: I retrieved the contents of each single sheet with help of regular expressions, put all records in one text-file, created an empty table in mysql, inserted the records from text-files into the mysql-tables and did some sql-queries for quality controls. And this I did all from ms-dos with Perl-programms. It's hand made but it works very well. Maybe more important than which programming philosophy (procedural, oo) is to keep the code clear, documented and well organised, but this may be obsolet to tell you.

By the way, if you want just learning sql without complex databases as MySQL you can use library sqldf in R or sqlite. sqlite is a sql-databases which is packed in only on file. There is an addon for Firefox which you can use to connect with sqlite and perform all your sql tasks: importing csv data etc.

+1 for sqlite, much easier with backups and no extra server running

Http://playingwithr.blogspot.com/2011/05/accessing-mysql-through-r.html

Enclosed is a MySQL Tutorial for your use.

http://downloads.mysql.com/docs/mysql-tutorial-excerpt-5.1-en.pdf

If you need a relational database, you usually end up needing a fuller implementation than MySQL provides. I would seriously suggest starting with Postgresql.

Many competent programmers find SQL hard to learn. It is a bit different to imperative programming languages - allow time for it to soak in.

PHP & redbeanphp (Orm) which connect mysql with php  would be great for a starter.