索引列以便在MySQL 5.6或更高版本中更快地进行查询

Question

I'm building a real estate app. I have a table called properties which is like the main table that has all common columns (10 columns) for all types of properties (lands, apartments, ... etc) and then I have a specific table for each property type since each type has some specific column. here is the property table:

CREATE TABLE `properties` (
  `property_id` int(11) NOT NULL AUTO_INCREMENT,
  `property_type` int(11) DEFAULT NULL,
  `property_title` varchar(255) NOT NULL,
  `property_description` varchar(1000) NOT NULL,
  `country_id` int(11) NOT NULL,
  `city_id` int(11) NOT NULL,
  `city_location_id` int(11) NOT NULL,
  `price` int(11) DEFAULT NULL,
  `area` decimal(7,2) DEFAULT NULL,
  `latitude` decimal(10,8) DEFAULT NULL,
  `longitude` decimal(11,8) DEFAULT NULL,
  `entry_date` datetime NOT NULL,
  `last_modification_date` datetime NOT NULL,
  PRIMARY KEY (`property_id`)
) 

and here is the apartments for example:

CREATE TABLE `apartments` (
  `apartment_id` INT NOT NULL COMMENT '',
  `num_of_bedrooms` INT NULL COMMENT '',
  `num_of_bathrooms` INT NULL COMMENT '',
  `num_of_garages` INT NULL COMMENT '',
  PRIMARY KEY (`apartment_id`)  COMMENT '',
  CONSTRAINT `properties_apartments_fk`
    FOREIGN KEY (`apartment_id`)
    REFERENCES `aqar_world`.`properties` (`property_id`)
    ON DELETE CASCADE
    ON UPDATE NO ACTION);

now the user can filter his search based on almost any of these columns or a combination of them, so how should I put my indexing strategy on the columns (the user could filter based on price, area, area and price, number of bedrooms and location and so on with these so many combinations) .. another point is that the property_description and property_title are texts so I'll have to add a fulltext index on each of them, right? also there is a join between these two tables and also between them and some other table (like agents tables for example).

I've read some say since mysql 5.6 there something in the optimizer that makes use of multiple indexes so you can put an index on each column but I don't know if that is right .. please advice since I'm not that good in taking care of DB performance

Answer 1

5.7 has JSON tricks. MariaDB 10 has Dynamic Columns with similar tricks.

The main principle: Expose the more useful fields; throw the more obscure fields into JSON or Dynamic columns. Then let MySQL filter on the former, and your app takes care of further filtering on the latter.

More discussion .