终于来学习目标 学习openGauss收集统计信息、打印执行计划、垃圾收集和checkpoint，统计信息是性能优化方面必不可少的，需要好好学习，掌握
1.连接数据库
Welcome to 墨天轮.
This is Web Terminal of modb.pro; Good Good Study, Day Day Up.
root@modb:~# su - omm
omm@modb:~$ gsql -r
gsql ((openGauss 2.0.0 build 78689da9) compiled at 2021-03-31 21:03:52 commit 0 last mr )
Non-SSL connection (SSL connection is recommended when requiring high-security)
Type "help" for help.
2.模拟数据
omm=# Create schema tpcds;
CREATE SCHEMA
omm=# CREATE TABLE tpcds.customer_address
omm-# (
omm(# ca_address_sk integer NOT NULL ,
omm(# ca_address_id character(16),
omm(# ca_street_number character(10) ,
omm(# ca_street_name character varying(60) ,
omm(# ca_street_type character(15) ,
omm(# ca_suite_number character(10) ,
omm(# ca_city character varying(60) ,
omm(# ca_county character varying(30) ,
omm(# ca_state character(2) ,
omm(# ca_zip character(10) ,
omm(# ca_country character varying(20) ,
omm(# ca_gmt_offset numeric(5,2) ,
omm(# ca_location_type character(20)
omm(# );
CREATE TABLE
omm=# insert into tpcds.customer_address values
omm-# (1, 'AAAAAAAABAAAAAAA', '18', 'Jackson', 'Parkway', 'Suite 280', 'Fairfield', 'Maricopa County', 'AZ', '86192' ,'United States', -7.00, 'condo'),
omm-# (2, 'AAAAAAAACAAAAAAA', '362', 'Washington 6th', 'RD', 'Suite 80', 'Fairview', 'Taos County', 'NM', '85709', 'United States', -7.00, 'condo'),
omm-# (3, 'AAAAAAAADAAAAAAA', '585', 'Dogwood Washington', 'Circle', 'Suite Q', 'Pleasant Valley', 'York County', 'PA', '12477', 'United States', -5.00, 'single family');
INSERT 0 3
omm=# omm=#
omm=# insert into tpcds.customer_address values(generate_series(10, 10000));
omm=# INSERT 0 9991
omm=# select relname, relpages, reltuples from pg_class where relname = 'customer_address';
relname | relpages | reltuples
------------------+----------+-----------
customer_address | 0 | 0
(1 row)
3.收集统计信息
omm=# analyze VERBOSE tpcds.customer_address;
INFO: analyzing "tpcds.customer_address"(gaussdb pid=1)
INFO: ANALYZE INFO : "customer_address": scanned 55 of 55 pages, containing 9994 live rows and 0 dead rows; 9994 rows in sample, 9994 estimated total rows(gaussdb pid=1)
ANALYZE
omm=# select relname, relpages, reltuples from pg_class where relname = 'customer_address';
relname | relpages | reltuples
------------------+----------+-----------
customer_address | 55 | 9994
(1 row)
4.执行分析
omm=# SET explain_perf_mode=normal;
omm=# SET
omm=# EXPLAIN SELECT * FROM tpcds.customer_address;
Seq Scan on customer_address (cost=0.00..154.94 rows=9994 width=151)
(1 row)
omm=# QUERY PLAN
-----------------------------------------------------------------------
omm=# EXPLAIN(FORMAT JSON) SELECT * FROM tpcds.customer_address;
QUERY PLAN
--------------------------------------------
[ +
{ +
"Plan": { +
"Node Type": "Seq Scan", +
"Relation Name": "customer_address",+
"Alias": "customer_address", +
"Startup Cost": 0.00, +
"Total Cost": 154.94, +
"Plan Rows": 9994, +
"Plan Width": 151 +
} +
} +
]
(1 row)
omm=# omm=#
omm=# EXPLAIN(COSTS FALSE)SELECT * FROM tpcds.customer_address;
QUERY PLAN
------------------------------
Seq Scan on customer_address
(1 row)
omm=# EXPLAIN SELECT SUM(ca_address_sk) FROM tpcds.customer_address WHERE ca_address_sk<100;
QUERY PLAN
-------------------------------------------------------------------------
Aggregate (cost=180.16..180.17 rows=1 width=12)
-> Seq Scan on customer_address (cost=0.00..179.93 rows=94 width=4)
Filter: (ca_address_sk < 100)
(3 rows)
5.创建索引
omm=# create index customer_address_idx on tpcds.customer_address(ca_address_sk);
CREATE INDEX
omm=# EXPLAIN SELECT * FROM tpcds.customer_address WHERE ca_address_sk<100;
QUERY PLAN
------------------------------------------------------------------------------------------------
[Bypass]
Index Scan using customer_address_idx on customer_address (cost=0.00..9.90 rows=94 width=151)
Index Cond: (ca_address_sk < 100)
(3 rows)
执行更新操作
omm=# update tpcds.customer_address set ca_address_sk = ca_address_sk + 1 where ca_address_sk <100;
UPDATE 93
回收垃圾
omm=# VACUUM (VERBOSE, ANALYZE) tpcds.customer_address;
INFO: vacuuming "tpcds.customer_address"(gaussdb pid=1)
INFO: index "customer_address_idx" now contains 10087 row versions in 31 pages(gaussdb pid=1)
DETAIL: 0 index row versions were removed.
0 index pages have been deleted, 0 are currently reusable.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
INFO: "customer_address": found 0 removable, 10087 nonremovable row versions in 55 out of 55 pages(gaussdb pid=1)
DETAIL: 93 dead row versions cannot be removed yet.
There were 0 unused item pointers.
0 pages are entirely empty.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
INFO: analyzing "tpcds.customer_address"(gaussdb pid=1)
INFO: ANALYZE INFO : "customer_address": scanned 55 of 55 pages, containing 9994 live rows and 93 dead rows; 9994 rows in sample, 9994 estimated total rows(gaussdb pid=1)
VACUUM
执行检查点
omm=# omm=# CHECKPOINT;
CHECKPOINT
6.删除数据
omm=# drop schema tpcds cascade;
NOTICE: drop cascades to table tpcds.customer_address
DROP SCHEMA
课后作业
1.创建分区表，并用generate_series(1,N)函数对表插入数据
omm=# Create schema alex;
CREATE SCHEMA
omm=# omm-# (
omm(# CREATE TABLE alex.sharding_tbl
ca_address_sk integer NOT NULL ,
omm(# ca_address_id character(16),
omm(# ca_street_number character(10) ,
omm(# ca_street_name character varying(60) ,
omm(# ca_street_type character(15) ,
omm(# ca_suite_number character(10) ,
omm(# omm(# ca_city character varying(60) ,
ca_county character varying(30) ,
omm(# ca_state character(2) ,
omm(# ca_zip character(10) ,
omm(# ca_country character varying(20) ,
omm(# ca_gmt_offset numeric(5,2) ,
omm(# ca_location_type character(20)
omm(# );
CREATE TABLE
omm=#
omm=#
omm=#
omm=# omm-# insert into alex.sharding_tbl values
omm-# (1, 'AAAAAAAABAAAAAAA', '18', 'Jackson', 'Parkway', 'Suite 280', 'Fairfield', 'Maricopa County', 'AZ', '86192' ,'United States', -7.00, 'condo'),
-7.00, 'condo'),AAAAA', '362', 'Washington 6th', 'RD', 'Suite 80', 'Fairview', 'Taos County', 'NM', '85709', 'United States', -
omm-# (3, 'AAAAAAAADAAAAAAA', '585', 'Dogwood Washington', 'Circle', 'Suite Q', 'Pleasant Valley', 'York County', 'PA', '12477', 'United States', -5.00, 'single family');
INSERT 0 3
omm=#
omm=#
omm=#
omm=#
omm=# insert into alex.sharding_tbl values(generate_series(10, 10000));
INSERT 0 9991
2.收集表统计信息
omm=# analyze VERBOSE alex.sharding_tbl ;
INFO: analyzing "alex.sharding_tbl"(gaussdb pid=1)
INFO: ANALYZE INFO : "sharding_tbl": scanned 55 of 55 pages, containing 9994 live rows and 0 dead rows; 9994 rows in sample, 9994 estimated total rows(gaussdb pid=1)
ANALYZE
3.显示简单查询的执行计划；建立索引并显示有索引条件的执行计划
omm=# EXPLAIN SELECT * FROM alex.sharding_tbl;
QUERY PLAN
-------------------------------------------------------------------
Seq Scan on sharding_tbl (cost=0.00..154.94 rows=9994 width=151)
(1 row)
4.更新表数据，并做垃圾收集
omm=# update alex.sharding_tbl set ca_address_sk = ca_address_sk + 1 where ca_address_sk <100;
UPDATE 93
omm=# VACUUM (VERBOSE, ANALYZE) alex.sharding_tbl;
INFO: vacuuming "alex.sharding_tbl"(gaussdb pid=1)
INFO: "sharding_tbl": found 0 removable, 10087 nonremovable row versions in 55 out of 55 pages(gaussdb pid=1)
DETAIL: 93 dead row versions cannot be removed yet.
There were 0 unused item pointers.
0 pages are entirely empty.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
INFO: analyzing "alex.sharding_tbl"(gaussdb pid=1)
INFO: ANALYZE INFO : "sharding_tbl": scanned 55 of 55 pages, containing 9994 live rows and 93 dead rows; 9994 rows in sample, 9994 estimated total rows(gaussdb pid=1)
VACUUM
5.清理数据
omm=# drop schema alex cascade;
NOTICE: drop cascades to table alex.sharding_tbl
DROP SCHEMA
