如何使用索引使Web应用程序中的数据库访问更快？

Question

In a web based java application , I have a table with this schema

 Column          Datatype

 Userid          char(25) primary key;
 name            char(100)
 address         varchar(120)
 email           char(50) UNIQUE;

with 120000 records. (Cardinality)

Now, I want to fetch the matching names (character by character) stored under column -> name. when I start typing in a search box (through ajax obviously).

eg if I write p then all names starting from letter p will be retrieved and would come in the list to populate.

Right now I am not using any indexing . So how could I use it to make the access/search faster .

The database I am using is MySql 5 and java as frontend .

Any suggestion for using appropriate collection for this purpose or any hashing would help in the database if yes please elaborate.

Though I have got the right track for this StackOverflowers. If I modify the question and use hashing in place of indexing with all its definition and usage statistics then how to do that and is it beneficial?

Answer 1

The simplest way is just:

ALTER TABLE `your_table` ADD INDEX(`name`);

And also, as names are variable length, use VARCHAR instead. For your key, use INT!

You can shrink your index to one byte if you search by one character only:

ALTER TABLE `your_table` ADD INDEX `name`(`name`(1));

Then, for better performance, you can use covering index. It means you need to have all required data indexed properly so it resides in memory (make sure you configured your server properly) and MySQL does not need to read from disk.

Answer 2

By default, PRIMARY KEY and UNIQUE properties are indexed in the table. So, your email and userid are already indexed. ( NOTE: I seriously suggest you not to use CHAR for userid )

And use

INDEX `<any-name-here>` (`name`) # The `name` is the column and `<any-
                                 # name-here>` is the name of index.

Answer 3

CREATE INDEX id_index ON employee_table (employeeID);

Answer 4

Create the indexing on the columns you will have in WHERE clause.

If your query will be most likely on name, create indexing on name column.

Remember this will leverage your insert/update record to that table which means your INSERT/UPDATE operations will be executed slower.

Answer 5

I wouldn't use DB queries. You could preload the (unique) names in a String array and perform a binary search for the characters typed so far. This ensures maximum responisveness.

Answer 6

In generell you have to understand how a database index works.

But for your problems, some suggestions:

If you are using something like this (sorry, I don't know the MySql wildcard character, assuming it is %)

select * from MyTable where name like '%a%'

where a is a letter the user entered - this will never work on bigger tables, because the first % says that there are any characters before a - this means the whole table has to be scanned (this means the index is nearly useless).

Consider implementing the Autocompleter so that it starts autocompleting after at least 3 characters, which would lead to SQL similar to this:

select * from MyTable where name like 'abc%'

This would allow the database to use the index, because it only has use the data beginning with "abc"

Unfortunately you do not have an index at all on the column which can be used for your query, so you can add it via the following command:

ALTER TABLE 'table' ADD INDEX('column_you_search')

Hope this helps.