SQL statement optimization steps (1)

        mysql Database is a common database processing language , It's also a limitation IT A database language widely used in the industry . Databases are used to process data , Add, delete, modify and check the data （CRUD）, For large capacity data processing , The events that need to be handled may be 1 God ,2 God , For the SQL The optimization of statements is very important .

        The following is from 5 There are three ways to understand SQL The basic steps of optimization

One 、 Query execution efficiency  

1、 Check the frequency of each command in the database

show global status like 'com_______'

2、 see innoDB Frequency of adding, deleting, modifying and querying

show global status like 'innoDB_rows_%' ;       

  Two 、 Locate inefficient execution SQL frequency

          The following two methods can be used to locate those with low execution efficiency SQL sentence

1、 Slow query log ：（ View the completed SQL How often statements are executed ）

2、show processlist ( View the running speed of statements in real time )

  The following describes the meaning of each field

title	meaning
Id	The sequence number of each statement previously executed allocated by the system
User	Current user name
Host	Display the host number sent by this statement 、 port , It can be used to track users who have problems
db	Displays which database this process is currently connected to
Command	Displays the execution status of the currently connected command , Sleep (sleep), Inquire about (query), Connect (connect)
Time	Shows the duration of this state
state	Show current SQL Statement state
Info	Which execution statement corresponds to the record of this row of data

Be careful ：state Status statement ：

            eg: select Statements are columns , To pass copying to tmp  table,sorting result,sending  data Wait for the status to complete .

3、 ... and 、explain Analysis execution plan

        stay sql Add... Before the sentence explain You can query the execution plan of this statement

explain select * from tb_score;

Analyze the meaning of each field ：

Field	meaning
id	select The serial number of the query , Indicates execution select Order of statements
select_type	Express select The type of , Common are SIMPLE( A simple watch )、PRIMARY( Primary key query )、UNION（ Link query ,union After statement ）、SUBQUERY( First in subquery select)
table	Output structure table
type	Indicates the type of table connection , Sort performance from high to low （system-->const-->eq_rf-->ref-->ref_or_null-->index_merge-->index_subquery-->range-->index-->all）
possible_keys	Indicates the index that may be used in query
key	Represents the index actually used in the query
key_len	Length of index used
rows	Number of scan lines
extra	Description and description of the implementation   using filesort  / using index/using tenporary

3.1  Id  select Query serial number

       d The fields are select The serial number of the query , It's a set of numbers , Represents execution in query select Clause or the order of the operation table , There is a concentrated situation

       1、 id The same indicates that the loading is smooth from top to bottom

       2、 id Different id The bigger the value is. , The higher the priority , Priority to be implemented

3.2 select_type Table type

        Value meaning ：（ From top to bottom, efficiency is getting lower and lower ）

simple ordinary select Inquire about , No subqueries or union

primary In a statement that contains subqueries , The outermost query is primary

subquery Subquery

derived （ derivative ）, Put the table of subquery in temporary table

union The joint query

union result from union Get the structure in the table select

3.3 table and type

         table Identify which table to source from

         type It shows the type of access , Is a more important indicator

          type , Access from the top down , The access speed is getting lower and lower

null Do not access any tables and indexes , Direct return select * from now();

system There is only one record in the table （ The system tables ）

const （ Constant unique index query ） It can be found by indexing only once , Generally used for primary keys and unique, There is only one record

eq_ref Associated query with only one result

ref Non unique index query , Returns all rows for a single value

range Range queries

index Traversed all the index numbers （b-tree structure ）

all Will traverse the entire table to find the matching rows

3.4  key

possible key Possible indexes

key Index actually used , If it is null I just didn't follow the index

key_len key Length of index , The shorter the better

3.5 rows

         rows Number of lines scanned

3.6 xtra

using filesort : Use files to sort , Low efficiency , Consumption performance

explain select * from student order by sname

here sname It's not an index , That is, you need to scan the whole disk and then sort , Low efficiency , The optimization process is to sname Create index .

using temporary Use a temporary watch , Low efficiency , Consumption performance

using index Use index , The fastest

Be careful ： If the first two appear, you need to consider optimization

Four 、show profile analysis sql

        testing sql Statement execution time

  4.1、 Check whether the current system supports profile Instructions

          select @@have_profiling;

  4.2、 Check profile Open or not

           select @@profiling;

   4.3、 Start profile function

           set @@profiling;

    4.4、 View the elapsed time of all previous instructions

            show profiles;

    4.5、 See where an instruction takes longer

               show prifile for query 5;          //query 5 Indicates the order of execution id

5、 ... and 、trace Analysis optimizer

sql sentence --> encapsulation -->trace Optimizer

1、 Start optimizer

 set optmizer_trace='enabled=on',end_markers_in_json=on;

2、 Set the occupied memory size

set optimizer_trace_max_men_size=1000000;

3、 View the optimizer execution plan

select  * from information_schma.optimizer_trace\G;