提高 Postgresql 中的函数性能

【问题标题】：Improve Function performance in Postgresql提高 Postgresql 中的函数性能
【发布时间】：2016-10-12 00:50:00
【问题描述】：

以下函数正在创建一个临时表，然后在其中填充数据。

该临时表应该在单击网站上的按钮时显示。

我尝试提高性能的步骤： 1.修改postgresql.conf文件。 2. 临时表增加索引

CREATE OR REPLACE FUNCTION lastonemonth()
  RETURNS void AS
$BODY$
  DECLARE
  query1 text;
  query2 text;
  var_loop1 RECORD;
  dealerName text;
  cur_minus_2_month text; 
  var_sum_of_quantity numeric;
  var_net_value numeric;
  var_average_price_nsr numeric; 
  BEGIN
  EXECUTE 'DROP TABLE IF EXISTS LOM CASCADE';
  EXECUTE 'CREATE TEMP TABLE LOM ( dealer_name text PRIMARY KEY , sum_of_quantity numeric, net_value numeric, average_price_nsr numeric)';
  EXECUTE 'CREATE UNIQUE INDEX dealer_name_idx ON LOM (dealer_name)';

  query1:= 'SELECT DISTINCT dealer FROM customernotorderdb_temp WHERE month IN ( to_char( now() - interval ''2 month'', ''YYYYMM'') , to_char( now() - interval ''3 month'', ''YYYYMM''))EXCEPT SELECT DISTINCT dealer FROM customernotorderdb_temp WHERE month IN ( to_char( now() - interval ''1 month'', ''YYYYMM''), to_char(now(), ''YYYYMM'') )';
  FOR var_loop1 IN EXECUTE (query1)
  LOOP
      dealerName:= var_loop1.dealer;
      INSERT INTO LOM(dealer_name) VALUES(dealerName);
      EXECUTE 'SELECT month FROM customernotorderdb_temp WHERE dealer LIKE ''%'||dealerName||'%'' AND month IN(to_char( now() - interval ''2 month'', ''YYYYMM''))' INTO cur_minus_2_month;

      --RAISE NOTICE 'cur_minus_2_month( % )', cur_minus_2_month;
      IF cur_minus_2_month IS NOT NULL 
      THEN
          EXECUTE 'SELECT SUM(saleqtypermt) FROM customernotorderdb_temp WHERE dealer LIKE ''%'||dealerName||'%'' AND month IN (to_char( now() - interval ''2 month'', ''YYYYMM''),to_char( now() - interval ''3 month'', ''YYYYMM''),to_char( now() - interval ''4 month'', ''YYYYMM''),to_char( now() - interval ''5 month'', ''YYYYMM''),to_char( now() - interval ''6 month'', ''YYYYMM''),to_char( now() - interval ''7 month'', ''YYYYMM''),to_char( now() -interval ''8 month'', ''YYYYMM''),to_char( now() - interval ''9 month'', ''YYYYMM''),to_char( now() - interval ''10 month'', ''YYYYMM''),to_char( now() - interval ''11 month'', ''YYYYMM''),to_char( now() - interval ''12 month'', ''YYYYMM''),to_char( now() - interval ''13 month'', ''YYYYMM''))' INTO var_sum_of_quantity;
          EXECUTE 'SELECT SUM(basic_value-rate_diff) FROM customernotorderdb_temp WHERE dealer LIKE ''%'||dealerName||'%'' AND month IN (to_char( now() - interval ''2 month'', ''YYYYMM''),to_char( now() - interval ''3 month'', ''YYYYMM''),to_char( now() - interval ''4 month'', ''YYYYMM''),to_char( now() - interval ''5 month'', ''YYYYMM''),to_char( now() - interval ''6 month'', ''YYYYMM''),to_char( now() - interval ''7 month'', ''YYYYMM''),to_char( now() -interval ''8 month'', ''YYYYMM''),to_char( now() - interval ''9 month'', ''YYYYMM''),to_char( now() - interval ''10 month'', ''YYYYMM''),to_char( now() - interval ''11 month'', ''YYYYMM''),to_char( now() - interval ''12 month'', ''YYYYMM''),to_char( now() - interval ''13 month'', ''YYYYMM''))' INTO var_net_value;
          EXECUTE 'SELECT SUM(avgpricensr) FROM customernotorderdb_temp WHERE dealer LIKE ''%'||dealerName||'%'' AND month IN (to_char( now() - interval ''2 month'', ''YYYYMM''),to_char( now() - interval ''3 month'', ''YYYYMM''),to_char( now() - interval ''4 month'', ''YYYYMM''),to_char( now() - interval ''5 month'', ''YYYYMM''),to_char( now() - interval ''6 month'', ''YYYYMM''),to_char( now() - interval ''7 month'', ''YYYYMM''),to_char( now() -interval ''8 month'', ''YYYYMM''),to_char( now() - interval ''9 month'', ''YYYYMM''),to_char( now() - interval ''10 month'', ''YYYYMM''),to_char( now() - interval ''11 month'', ''YYYYMM''),to_char( now() - interval ''12 month'', ''YYYYMM''),to_char( now() - interval ''13 month'', ''YYYYMM''))' INTO var_average_price_nsr;
          --RAISE NOTICE 'A [ %   % ]', dealername,var_net_value;
          UPDATE LOM SET sum_of_quantity=var_sum_of_quantity,
                         net_value=var_net_value,
                         average_price_nsr=var_average_price_nsr 
          WHERE dealer_name=var_loop1.dealer;

      ELSE
          EXECUTE 'SELECT SUM(saleqtypermt) FROM customernotorderdb_temp WHERE dealer LIKE ''%'||dealerName||'%'' AND month IN (to_char( now() - interval ''3 month'', ''YYYYMM''),to_char( now() - interval ''4 month'', ''YYYYMM''),to_char( now() - interval ''5 month'', ''YYYYMM''),to_char( now() - interval ''6 month'', ''YYYYMM''),to_char( now() - interval ''7 month'', ''YYYYMM''),to_char( now() -interval ''8 month'', ''YYYYMM''),to_char( now() - interval ''9 month'', ''YYYYMM''),to_char( now() - interval ''10 month'', ''YYYYMM''),to_char( now() - interval ''11 month'', ''YYYYMM''),to_char( now() - interval ''12 month'', ''YYYYMM''),to_char( now() - interval ''13 month'', ''YYYYMM''),to_char( now() - interval ''14 month'', ''YYYYMM''))' INTO var_sum_of_quantity;
          EXECUTE 'SELECT SUM(basic_value-rate_diff) FROM customernotorderdb_temp WHERE dealer LIKE ''%'||dealerName||'%'' AND month IN (to_char( now() - interval ''3 month'', ''YYYYMM''),to_char( now() - interval ''4 month'', ''YYYYMM''),to_char( now() - interval ''5 month'', ''YYYYMM''),to_char( now() - interval ''6 month'', ''YYYYMM''),to_char( now() - interval ''7 month'', ''YYYYMM''),to_char( now() -interval ''8 month'', ''YYYYMM''),to_char( now() - interval ''9 month'', ''YYYYMM''),to_char( now() - interval ''10 month'', ''YYYYMM''),to_char( now() - interval ''11 month'', ''YYYYMM''),to_char( now() - interval ''12 month'', ''YYYYMM''),to_char( now() - interval ''13 month'', ''YYYYMM''),to_char( now() - interval ''14 month'', ''YYYYMM''))' INTO var_net_value;
          EXECUTE 'SELECT SUM(avgpricensr) FROM customernotorderdb_temp WHERE dealer LIKE ''%'||dealerName||'%'' AND month IN (to_char( now() - interval ''3 month'', ''YYYYMM''),to_char( now() - interval ''4 month'', ''YYYYMM''),to_char( now() - interval ''5 month'', ''YYYYMM''),to_char( now() - interval ''6 month'', ''YYYYMM''),to_char( now() - interval ''7 month'', ''YYYYMM''),to_char( now() -interval ''8 month'', ''YYYYMM''),to_char( now() - interval ''9 month'', ''YYYYMM''),to_char( now() - interval ''10 month'', ''YYYYMM''),to_char( now() - interval ''11 month'', ''YYYYMM''),to_char( now() - interval ''12 month'', ''YYYYMM''),to_char( now() - interval ''13 month'', ''YYYYMM''),to_char( now() - interval ''14 month'', ''YYYYMM''))' INTO var_average_price_nsr;
          --RAISE NOTICE 'B [ %  % ]', dealername,var_net_value;
          UPDATE LOM SET sum_of_quantity=var_sum_of_quantity,net_value=var_net_value,average_price_nsr=var_average_price_nsr WHERE dealer_name=var_loop1.dealer;
      END IF;         
 END LOOP;
END;
$BODY$
  LANGUAGE plpgsql VOLATILE

【问题讨论】：

  • 您没有在表格上运行 ANALYZE。较新的 drop temp 表，改为截断它。然后就可以使用静态SQL了。
  • 你不需要函数；你不需要循环，你不需要 plpgsql。该函数有效地执行INSERT INTO LOM(...) SELECT dealer_name, SUM(xxx0), ... FROM customernotorderdb_temp GROUP BY dealer_name（以及其他一些聚合和选择）
  • 经过上述建议后，代码如下：
  • WHERE dealer LIKE ''%'||dealerName||'%'' 为什么在这里使用通配符？你从同一张桌子上得到了经销商的名字！

标签： sql performance postgresql function


【解决方案1】：

首先，您似乎将月份存储为字符串并手动与字符串进行日期比较。那会让它慢下来。你能用 Postgres 日期类型或整数来代替吗？

对于其他事情，如果不知道您的数据库布局和统计信息，我真的无法判断，但这里是优化查询的基本方法：

  1. 使用 psql 或其他工具打开新连接。
  2. set enable_seqscan = off，仅在此会话中禁用查询计划器中的 seq 扫描（除非完全必要）。
  3. 在查询之前运行EXPLAIN 以查看哪些查询被强制扫描大表。如果您看到任何 SEQ SCANs，请添加索引。

【讨论】：

  • "EXPLAIN ANALYZE SELECT lastonemonth();st=0.00..0.26 rows=1 width=0) (实际时间=16959.720..16959.721 rows=1 loops=1)" "总运行时间：16959.734毫秒”
  • 非常感谢，现在代码的效率是 EXPLAIN ANALYZE SELECT lastonemonth();"Result (cost=0.00..0.26 rows=1 width=0) (实际时间=2783.845..2783.845 rows=1 loops=1)" "总运行时间：2783.859 毫秒"
  • 不错。你改变了什么？您没有发布 EXPLAIN 生成的查询计划。
  • 1.删除通配符 % 2. 添加显式索引。 3. if 块下将三个execute 语句合并为一个 4. 修改postgresql.conf 文件设置 5. 从函数中删除'CREATE' 和'DROP' 语句 6. 删除减速部分的一个变量
猜你喜欢
  • 1970-01-01
  • 1970-01-01
  • 2013-12-17
  • 2011-09-30
  • 2015-07-11
  • 2014-03-20
  • 2011-01-06
  • 2021-08-03
  • 2013-02-17
相关资源
最近更新 更多
热门标签
Java Python linux javascript Mysql C# Docker 算法 前端 SpringBoot Redis Vue spring 设计模式 .net core .net kubernetes c++ 数据库 数据结构 大数据 js 机器学习 微服务 Android Go 程序员 面试 JVM ASP.net core 云原生 人工智能 后端 PHP git CSS golang k8s Nginx Django mybatis 深度学习 多线程 React 架构 devops 爬虫 云计算 Spring Boot LeetCode