Retired Note 1 # MySQL = lambdasql: SQL + & amp; #39; SourceCode4ExplainPlan & amp; #39;

Retired Note 1 # MySQL = lambdasql: SQL + & #39; SourceCode4ExplainPlan & #39;

The Mysql query execution process is roughly divided into four stages:

Syntax analysis (SQL _parse.cc <lexical analysis, syntax analysis, semantic check>)

> SQL _resolver.cc # JOIN. prepare

Generate logical query plan (SQL _optimizer.cc)

>># JOIN. optimize

Generate a physical query plan (SQL _planner.cc)

Run the explain plan (SQL _executor.cc) JOIN.exe c

JOIN. prepare ():

Remove redundant sub-queries

In subquery optimization resolve_subquery ()

Converts all/any subqueries to min/max to optimize simple subqueries.

JOIN. optimize ():

Pull flatten_subqueries () on the subquery ()

Optimize external connection to internal connection simplify_joins ()

Eliminate nested connections

Where clause, join/on clause, having clause simplification, expression simplification containing constants, equality merge optimize_cond () opt_sum_query ()

Optimize count (*) min () man () without the group by clause ()

Determine the multi-Table connection path make_join_statistice ()

Optimize distinct

Create temporary table store temporary results optimize group sorting choose_table_order ()

In fact, logical and physical optimization is a bit confusing and ambiguous,

After the physical query is optimized, some logic optimization is continued.

There are only two ways to scan a single table:

Index scan (others), table scan (JT_ALL)

Opt_range.h> class QUICK_SELECT_ I> use the index scanning drop-down Interface

For the child classes, see source code;

Main classes and struct:

Query tree class st_select_lex

Index struct st_key_create_information class Key_part_spec

Connection table st_join_table

JOIN: public SQL _alloc

Condition COND_EQUAL

Location struct st_position

Cost Estimation Cost_estimate

SELECT NOW(),CURDATE(),CURTIME()