Right approach to avoid duplicate rows

Question

This for private social network, so I wonder,

Which approach is more accurate and valid at this time. So I estimate the tables to have million rows,

First:

CREATE TABLE `system_auth_users` (

  `username` char(15) NOT NULL PRIMARY KEY,
  `password` varchar(32) NOT NULL,
  INDEX `username` (`username`)

)type=MyISAM;


CREATE TABLE `system_auth_info` (

  `username` char(15) NOT NULL PRIMARY KEY,
  `fname`    varchar(15) NOT NULL,
  `lname`    varchar(15) NOT NULL,
  `gender`    char(1) NOT NULL,
  `location`  varchar(3) DEFAULT NULL, 
  `descr`     varchar(15) DEFAULT NULL,
   INDEX `username` (`username`)

)type=MyISAM DEFAULT CHARSET=UTF8;


CREATE TABLE `system_auth_pass_recovery` (

  `username` char(15) NOT NULL PRIMARY KEY,
  `token`   varchar(35) NOT NULL,
  `time` int(10) NOT NULL   
   INDEX `username` (`username`)

)type=MyISAM DEFAULT CHARSET=UTF8;

The problem is that this approach is really good for maintenance/for queries, but it wastes free space by using duplicate row username

2 :

Second approach is to create single table for all common tasks (authorization, password recovery, profile)

CREATE TABLE `system_auth_info` (

  `username` char(15) NOT NULL PRIMARY KEY, /* for password recovery, for authorization, for 'personal info' */
  `password` varchar(32) NOT NULL,
  `token`   varchar(35) NOT NULL,
  `fname`    varchar(15) NOT NULL,
  `lname`    varchar(15) NOT NULL,
  `gender`    char(1) NOT NULL,
  `location`  varchar(3) DEFAULT NULL, 
  `descr`     varchar(15) DEFAULT NULL,
   INDEX `username` (`username`)

)type=MyISAM DEFAULT CHARSET=UTF8;

the benefit of this approach is that this one doesnt waste memory, but makes a little bit harder to maintain whole system

So, the question is, Which is approach is recommended for >= 1 000 000 rows ( including perfomance ) ?

Thanks.

Answer 1

So after spending days and night, I found the answer, The first approach is far better for common purpose. Twitter and Facebook use that one