Question

I am collecting data and storing it MySQL, for:

• 75 variables
• 55 countries
• Each year

I have, at this stage since I am building this tool created a single table, of variables / countries (storing 1 year worth of data).

Next year (and for several years after that) a new set of data will be input for each country.

There are therefore 3 variables in controlling data returned to a user reviewing all collected data. The general form of any query would be:

• Show me these specifics variables, for these specific countries, for these specific years.

(Show me average age and weight, for USA and Canada, for 2012 and 2009, for example)

My question is, it seems that I have two options for arranging this data:
-Multiple tables where I create a table of country / variable for each year data is collected
– Single table and simply add a column (field) for the year that data relates to.

As far as I can tell I could make these database calls with either sructure, but is one more powerful / efficient / quicker, and why?

Thanks for your consideration.

It’s a PDO / PHP interface if that is relevent.

Answer 1

Using a relational approach generally involves more tables. This translates into queries being a bit more slow (though probably not noticeable in small databases) and database size to be smaller. This makes it simpler to update information properly and thus ensure data integrity. For example, if Joe’s address changes you know it will be changed on all reports using Joe’s address.

Using less linked tables where one field can be repeated multiple times you risk having disparity between data from different tables where you would naturally expect it to be equal. Access speed should be a bit faster if you arrange your tables properly because your information will be grouped according to how you access it.

For example, in the first method you would have an Orders table with a Supplier and Client table to make a complete invoice whereas in the second method you would want to put some information of both Supplier and Client in the Orders table such that accessing that finding the row corresponding to the invoice number you are looking for would return the entire set of data that you need (thus eliminating the need for joins on Supplier and Client and reducing load on the database server).

Edit: I think a better answer would require a bit more information about your data (samples for example).