两个子表的MySQL JOIN问题

Question

I'm learning MySQL JOINs now and one thing I want to know is how to JOIN two child tables with one father table.

In this case, the query is this:

SET @OLD_UNIQUE_CHECKS=@@UNIQUE_CHECKS, UNIQUE_CHECKS=0;
SET @OLD_FOREIGN_KEY_CHECKS=@@FOREIGN_KEY_CHECKS, FOREIGN_KEY_CHECKS=0;
SET @OLD_SQL_MODE=@@SQL_MODE, SQL_MODE='TRADITIONAL';

DROP SCHEMA IF EXISTS `test` ;

CREATE SCHEMA IF NOT EXISTS `test` DEFAULT CHARACTER SET utf8 COLLATE utf8_general_ci ;

USE `test`;

CREATE  TABLE IF NOT EXISTS `test`.`objetos` (
  `idobjetos` INT(11) NOT NULL AUTO_INCREMENT ,
  `modelo` VARCHAR(45) NULL DEFAULT NULL ,
  `descricao` VARCHAR(45) NULL DEFAULT NULL ,
  `token_dono` VARCHAR(41) NOT NULL ,
  PRIMARY KEY (`idobjetos`, `token_dono`) ,
  INDEX `token_dono` (`token_dono` ASC) )
ENGINE = InnoDB
DEFAULT CHARACTER SET = utf8
COLLATE = utf8_general_ci;

CREATE  TABLE IF NOT EXISTS `test`.`empresas` (
  `idempresa` INT(11) NULL DEFAULT NULL AUTO_INCREMENT ,
  `nome` VARCHAR(45) NULL DEFAULT NULL ,
  `cnpj` VARCHAR(45) NULL DEFAULT NULL ,
  `email` VARCHAR(45) NULL DEFAULT NULL ,
  `telefone` VARCHAR(45) NULL DEFAULT NULL ,
  `token` VARCHAR(41) NOT NULL ,
  PRIMARY KEY (`idempresa`, `token`) ,
  INDEX `token_empresa` (`token` ASC) ,
  UNIQUE INDEX `cnpj_UNIQUE` (`cnpj` ASC) ,
  UNIQUE INDEX `email_UNIQUE` (`email` ASC) ,
  UNIQUE INDEX `telefone_UNIQUE` (`telefone` ASC) ,
  CONSTRAINT `token_empresa`
    FOREIGN KEY (`token` )
    REFERENCES `test`.`objetos` (`token_dono` )
    ON DELETE NO ACTION
    ON UPDATE NO ACTION)
ENGINE = InnoDB
DEFAULT CHARACTER SET = utf8
COLLATE = utf8_general_ci;

CREATE  TABLE IF NOT EXISTS `test`.`civis` (
  `idcivil` INT(11) NOT NULL AUTO_INCREMENT ,
  `nome` VARCHAR(45) NULL DEFAULT NULL ,
  `cpf` VARCHAR(45) NULL DEFAULT NULL ,
  `email` VARCHAR(45) NULL DEFAULT NULL ,
  `token` VARCHAR(41) NOT NULL ,
  PRIMARY KEY (`idcivil`, `token`) ,
  INDEX `token_civil` (`token` ASC) ,
  UNIQUE INDEX `cpf_UNIQUE` (`cpf` ASC) ,
  UNIQUE INDEX `email_UNIQUE` (`email` ASC) ,
  CONSTRAINT `token_civil`
    FOREIGN KEY (`token` )
    REFERENCES `test`.`objetos` (`token_dono` )
    ON DELETE NO ACTION
    ON UPDATE NO ACTION)
ENGINE = InnoDB
DEFAULT CHARACTER SET = utf8
COLLATE = utf8_general_ci;

CREATE  TABLE IF NOT EXISTS `test`.`an_users` (
  `id` INT(11) NULL DEFAULT NULL AUTO_INCREMENT ,
  `usuario` VARCHAR(45) NULL DEFAULT NULL ,
  `senha` VARCHAR(45) NULL DEFAULT NULL ,
  `cod_usuario` INT(11) NOT NULL ,
  PRIMARY KEY (`id`, `cod_usuario`) ,
  INDEX `cod_usuario_fk` (`cod_usuario` ASC) ,
  CONSTRAINT `cod_usuario_fk`
    FOREIGN KEY (`cod_usuario` )
    REFERENCES `test`.`empresas` (`idempresa` )
    ON DELETE NO ACTION
    ON UPDATE NO ACTION)
ENGINE = InnoDB
DEFAULT CHARACTER SET = utf8
COLLATE = utf8_general_ci;


SET SQL_MODE=@OLD_SQL_MODE;
SET FOREIGN_KEY_CHECKS=@OLD_FOREIGN_KEY_CHECKS;
SET UNIQUE_CHECKS=@OLD_UNIQUE_CHECKS;

And I've inserted these values:

INSERT INTO `mydb`.`objetos` VALUES (NULL, 'Cadeira XTT', 'Quê?', SHA1(CONCAT(3913123612,'one@gmail.com')));
INSERT INTO `mydb`.`objetos` VALUES (NULL, 'Mesa TTX', 'Hein?', SHA1(CONCAT(4313123612,'two@gmail.com')));
INSERT INTO `mydb`.`objetos` VALUES (NULL, 'Prédio TT', 'Hein?', SHA1(CONCAT(73358847000116,'buzz@gmail.com')));

INSERT INTO `mydb`.`civis` VALUES (NULL, 'One', 3913123612, 'one@gmail.com', SHA1(CONCAT(3913123612,'one@gmail.com')));
INSERT INTO `mydb`.`civis` VALUES (NULL, 'Two', 4313123612, 'two@gmail.com', SHA1(CONCAT(4313123612,'two@gmail.com')));

INSERT INTO `mydb`.`empresas` VALUES(NULL, 'Buzz', 73358847000116, 'buzz@gmail.com', 33270743, SHA1(CONCAT(73358847000116,'buzz@gmail.com')));

But there are some problems when I want to retrieve the recorded data. The NATURAL JOIN doesn't work as the expected, taking only the rows with the same token and I can't to this:

SELECT * 
FROM objetos 

JOIN civis 
ON(objetos.token_dono = civis.token) 

JOIN empresas 
ON(objetos.token_dono = empresas.token)

To list all "objetos" from database with correct informations. But as I've tested it not work.

If someone can give me a light with these problems I'll be grateful.

Answer 1

The situation you are experiencing is that the unqualified JOIN is actually interpreted as INNER JOIN see the documentation for more details.

So when you run your query the row not containing a matching token from objetos is dropped and then since no token from civis match token from empresas all rows are dropped.

So from what I gather you are trying to do you will actually need to do a LEFT JOIN if you want to get all columns from all tables or separate this one query into to like so:

SELECT * FROM objetos o JOIN civis c ON (o.token_dono = c.token);

and

SELECT * FROM objetos o JOIN empresas e ON (o.token_dono = e.token);