Designing a database - Games played table - database

Question

I have pretty basic question. I'm designing a database to keep track of the chess games being played among certain group of individuals. I am thinking of creating one player table which will hold the details of each individual player. The primary key would be player ID.
My question is about associating games with players.
My initial thought is to create a games table. Each record in it will be a game, it will have 2 columns for players(player ids) who played it and third column will be a winner( or draw ).
What is the best way to design it? Should third columns value be a player id or I can just specify 1 if player in first column has won the game otherwise 0. How do I handle draw case? Do I insert some keyword?
I'd also like to know if there is some completely other way of designing this database.

Answer 1

Sublime, there are several ways you could define the structure of this database. Which design you ultimately choose depends on the performance expectations and the ultimate maintenance requirements of your database. I would recommend reading up on normalization which will give you some insights into the pros and cons of designing your tables in various ways.
To answer your specific question, I would:
Make a games table that consisted of the fields:
Game ID (primary key)
Player 1 ID (Foreign key to your players table)
Player 2 ID (Foreign key to your players table)
Game result
Important points to note here:
I included a game id primary key to uniquely identify one game from another. Without this, you will not be able to differentiate multiple games played by the same players. You may or may not require this in your implementation.
Game result could contain different sets of values based on your method of extracting the data into some meaningful representation. For example, some possibilities are to:
Have game result equal player id or 0 or -1 to indicate a draw. Least preferred, in my opinion, since this results in a consistency problem in the data that the field represents.
Have game result equal a pre-defined list of possible values (1 = Player 1 won, 2 = Player 2 won, 3 = Draw, 4 = Still in progress, etc). The possible values could be it's own table. This approach requires more extensive query planning.
You could also structure your data to consist of:
A players table - Player information
A games table - Game information
A results table - Game results information
In this setup, your games table would include only game id and the players id's. The results table would include the game id and the result of the game using one of the above solutions for identifying the winner.

Answer 2

Here are my views:
I don't like the idea of having the two columns 'player 1' and 'player 2'. When trying to get a list of all games 'Jeff' has played, you will have to check both 'player 1' and 'player 2' columns for his Id.
You could fix the above point by implementing a 'GamePlay' type of table, which has separate rows for each player. This way each player has their own outcome and are are held in a single column for easy querying.
A separate table for for possible outcomes would be tidy.
I've drafted up an example below
GameResult - Reference table for game results
GameResultId | Desc
01 | Won
02 | Lost
03 | Drew
Game - This is the game table, lists games. Can have details of location, date, time ect.
GameId | StartTime | Location | Ect
01 | 13:00 | Park | ect..
GamePlay - Have a row for each member-game interaction.
GamePlayId | GameId | PlayerId | GameResultId
01 | 01 | 01 | 01
02 | 01 | 02 | 02
This way, you can query by your players a lot easier.
(Such as getting a list of all players who have won a match)
SELECT
p.PlayerName
FROM
Player p
INNER JOIN GapePlay gp
ON p.PlayerId = gp.PlayerId
WHERE gp.GameresultId = 1
You could still easy query by game too
SELECT
gp.GameId
, p.Name
, gr.Desc
FROM
GamePlay gp
INNER JOIN Player p
ON gp.PlayerId = p.PlayerId
INNER JOIN GameResult gr
ON gp.GameresultId = gr.GameResultId
WHERE
GamePlayId = 01