PostgreSQL代码分析，查询优化部分，pull_ands()和p

PostgreSQL代码分析，查询优化部分，pull_ands()和pull_ors()，pull_andspull_ors

PostgreSQL代码分析，查询优化部分。

这里把规范谓词表达式的部分就整理完了，阅读的顺序如下：

一、PostgreSQL代码分析，查询优化部分，canonicalize_qual

二、PostgreSQL代码分析，查询优化部分，pull_ands()和pull_ors()

三、PostgreSQL代码分析，查询优化部分，process_duplicate_ors

*************************************************************************************************************************************************************

pull_ands()和pull_ors()的代码比较便于理解，就是把树状结构的AND操作拉平，下图是pull_ands的例子，pull_ors逻辑相同：

/*
 * pull_ands
 *	  Recursively flatten nested AND clauses into a single and-clause list.
 *
 * Input is the arglist of an AND clause.
 * Returns the rebuilt arglist (note original list structure is not touched).
 */
static List *
pull_ands(List *andlist)
{
	List	   *out_list = NIL;
	ListCell   *arg;

	foreach(arg, andlist)
	{
		Node	   *subexpr = (Node *) lfirst(arg);

		/*
		 * Note: we can destructively concat the subexpression's arglist
		 * because we know the recursive invocation of pull_ands will have
		 * built a new arglist not shared with any other expr. Otherwise we'd
		 * need a list_copy here.
		 */
		if (and_clause(subexpr))
			out_list = list_concat(out_list,
								   pull_ands(((BoolExpr *) subexpr)->args));
		else
			out_list = lappend(out_list, subexpr);
	}
	return out_list;
}

/*
 * pull_ors
 *	  Recursively flatten nested OR clauses into a single or-clause list.
 *
 * Input is the arglist of an OR clause.
 * Returns the rebuilt arglist (note original list structure is not touched).
 */
static List *
pull_ors(List *orlist)
{
	List	   *out_list = NIL;
	ListCell   *arg;

	foreach(arg, orlist)
	{
		Node	   *subexpr = (Node *) lfirst(arg);

		/*
		 * Note: we can destructively concat the subexpression's arglist
		 * because we know the recursive invocation of pull_ors will have
		 * built a new arglist not shared with any other expr. Otherwise we'd
		 * need a list_copy here.
		 */
		if (or_clause(subexpr))
			out_list = list_concat(out_list,
								   pull_ors(((BoolExpr *) subexpr)->args));
		else
			out_list = lappend(out_list, subexpr);
	}
	return out_list;
}

张大明白的blog：http://blog.csdn.net/shujiezhang

更新第一条数据的时候

1+1 = 2

该列上 2 已经存在，所以 update 会出错。

一次 update 有唯一性限制的列的多条记录，可不能这样干。
 

下面演示下：一 方法一：通过 pg_user 视图查询--1.1 设置用户的 log_statement 参数 postgres=# alter role francs set log_statement="all";ALTER ROLE --1.2 验证 postgres=# select * From pg_user where usename='francs';-[ RECORD 1 ]--------------------usename | francsusesysid | 24920usecreatedb | fusesuper | fusecatupd | fuserepl | fpasswd | ********valuntil | useconfig | {log_statement=all}--1.3 设置用户的 maintenance_work_mem 参数 postgres=# alter role francs set maintenance_work_mem="1GB";ALTER ROLE--1.4 再次验证 postgres=# select * From pg_user where usename='francs';-[ RECORD 1 ]---------------------------------------------usename | francsusesysid | 24920usecreatedb | fusesuper | fusecatupd | fuserepl | fpasswd | ********valuntil | useconfig | {log_statement=all,maintenance_work_mem=1GB} 备注：上面是通过 pg_user.useconfig 查询。二 方法二： 通过 pg_db_role_setting catalog 基表查询--2.1 pg_db_role_setting 表结构 Table "pg_catalog.pg_db_role_setting" Column | Type | Modifiers -------------+--------+----------- setdatabase | oid | not null setrole | oid | not null setconfig | text[] | Indexes: "pg_db_role_setting_databaseid_rol_index" UNIQUE, btree (setdatabase, setrole), tablespace "pg_global"备注：可见 pg_db_role_setting 会针对数据库，用户级别进行记录。--2.2 验证 postgres=# select oid,rolname from pg_authid where rolname='francs'; oid | rolname -------+--------- 24920 | francs(1 row)postgres=# select * From pg_db_role_setting where setrole=24920; setdatabase | setrole | setconfig -----------......余下全文>>