In this post, I am sharing necessary steps to remove or uninstall PostgreSQL 9.6 from Ubuntu OS. Yesterday, I was testing a few latest extensions in my PostgreSQL test server, and unfortunately, PostgreSQL Server corrupted.

As this is my test server, so simply I uninstalled full setup and again installed it.

PostgreSQL: How to Install PostgreSQL 9.6 on Ubuntu 16.04?

Please check below commands which I applied to uninstalled PostgreSQL:

First check Postgres related folders:

ubuntu@ip:~/anvesh$ dpkg -l | grep postgres

Result:

ii  pgdg-keyring               2014.1                  all     keyring for apt.postgresql.org
ii  postgresql-9.6             9.6.1-2.pgdg14.04+1     amd64   object-relational SQL database, version 9.6 server
ii  postgresql-client-9.6      9.6.1-2.pgdg14.04+1     amd64   front-end programs for PostgreSQL 9.6
ii  postgresql-client-common   178.pgdg14.04+1         all     manager for multiple PostgreSQL client versions
ii  postgresql-common          178.pgdg14.04+1         all     PostgreSQL database-cluster manager
ii  postgresql-contrib-9.6     9.6.1-2.pgdg14.04+1     amd64   additional facilities for PostgreSQL

Command to remove few postgres folders:

ubuntu@ip:~/anvesh$ sudo apt-get --purge remove postgresql postgresql-doc

Now check again Postgres related folders:

ubuntu@ip:~/anvesh$ dpkg -l | grep postgres

Result:

ii  pgdg-keyring              2014.1                all    keyring for apt.postgresql.org
ii  postgresql-client-9.6     9.6.1-2.pgdg14.04+1   amd64  front-end programs for PostgreSQL 9.6
ii  postgresql-client-common  178.pgdg14.04+1       all    manager for multiple PostgreSQL client versions

Now remove all other pending folders:

ubuntu@ip:~/anvesh$ sudo apt-get --purge remove postgresql postgresql-client-9.6 postgresql-common pgdg-keyring

Now PostgreSQL is uninstalled so confirmt that PSQL is running or not:

ubuntu@ip:~/anvesh$ psql
The program 'psql' is currently not installed. You can install it by typing:
sudo apt install postgresql-client-common

Now check the Postgres service status:

ubuntu@ip:~/anvesh$ sudo service postgresql status
● postgresql.service
   Loaded: not-found (Reason: No such file or directory)
   Active: inactive (dead)

In this post, I am sharing a new feature CROSSTABVIEW in PSQL of PostgreSQL 9.6. Many times we require data arrangement horizontally, instead of vertically.

Something like we are doing in a spreadsheet where one element has a multiple child elements and require calculation horizontally for each group of items.

Now, you can use CROSSTABVIEW in new PSQL for this kind of requirement.

Below is a small demonstration of this:

Open a psql command line and create a below sample table:

create table tbl_range as
select
    ('{a,b,c,d,e,f}'::text[])[1 + floor(random() * 6)] as data1
    ,('{g,h,i,j,k,l}'::text[])[1 + floor(random() * 6)] as data2
    ,random() * 1000 as randomnumber
from generate_series(1,10000);

Execute below few SELECTs:

select * from tbl_range;
select data1, data2, count(*) from tbl_range group by data1, data2;

Now check sample SELECT with CROSSTABVIEW:

select data1, data2, count(*) from tbl_range group by data1, data2 \crosstabview data1 data2 count;

The result:

dbrnd=# select data1, data2, count(*) from tbl_range group by data1, data2 \crosstabview data1 data2 count;
 data1 |  k  |  j  |  i  |  h  |  g  |  l
-------+-----+-----+-----+-----+-----+-----
 f     | 281 | 273 | 261 | 265 | 268 | 258
 a     | 288 | 293 | 278 | 293 | 248 | 279
 d     | 293 | 264 | 278 | 258 | 289 | 291
 c     | 272 | 276 | 292 | 287 | 263 | 281
 e     | 269 | 279 | 281 | 257 | 274 | 318
 b     | 289 | 270 | 301 | 295 | 247 | 291
(6 rows)

If you don’t know what VM and FSM is, I would suggest to read below articles:

PostgreSQL: What is Visibility Map (VM)?

PostgreSQL: What is a Free Space Map (FSM)?

PostgreSQL 9.6 introduced a pg_visibility module to get the visibility map information of your table, and it is also giving page level information.
For removing fragmentation, and improve the query performance we should check the visibility map of our table.

For that, you need to create a pg_visibility module in PostgreSQL 9.6 and use it.
Previously, people are checking the suffix _vm files which associate with your relation.

Check the below demonstration on how to use pg_visibility:

Create a pg_visibility extension:

CREATE EXTENSION pg_visibility;

Create a sample table with 10000 dummy records:

CREATE TABLE tbl_checkvmap AS SELECT generate_series(1,10000) AS id;

Check the visibility map:

SELECT * FROM pg_visibility('tbl_checkvmap'::regclass);

PostgreSQL 9.6 introduced new two columns of table pg_stat_activity. The columns are wait_event_type and wait_event.

The column wait_event_type reports the type of event a backend is waiting for, and the wait_event which is the name of the event being waiting for. These both columns are very useful for finding a query event information.

Please check small demonstration of this:

Create a sample table:

CREATE TABLE tbl_students(rno int, name character varying(10));

Put your table in lock mode:

BEGIN TRANSACTION;
LOCK tbl_students IN ACCESS EXCLUSIVE MODE;

Open a new query session, and try to insert a below record:

INSERT INTO tbl_students VALUES (1,'Anvesh');

Open a new query session, and check the output of pg_stat_activity:

SELECT query, wait_event_type, wait_event FROM pg_stat_activity

PostgreSQL 9.6 came with lots of very useful features and new system views.
I shared many articles on PostgreSQL 9.6 which you can get in PostgreSQL category.

PostgreSQL 9.6 also introduced a new system view which is pg_config. Using this view, we can get the information like important paths, postgres compilation details.

Query to pg_config:

select name, setting from pg_config;

The example result is:

PostgreSQL 9.6 introduced a pg_blocking_pids() to get a blocking process id of your running transaction.

When multiple sessions are trying to access a table, and that table is locked. At that time, we need to find that blocking process id which blocked that table so that we can take necessary steps.

Using pg_blocking_pids(), we can get that head process id which is blocking other connections.

Below is a full demonstration of this:

First, save your current process id:

SELECT pg_backend_pid();
-- my process id is: 6720

Create a sample table:

CREATE TABLE tbl_students(rno int, name character varying(10));

Put your table in lock mode:

BEGIN TRANSACTION;
LOCK tbl_students IN ACCESS EXCLUSIVE MODE;

Open a new session, and save process id for that session also:

SELECT pg_backend_pid();
-- my process id is: 5936

Now in that new session, try to insert below record:
Your result will be in waiting mode…

INSERT INTO tbl_students VALUES (1,'Anvesh');

Open a third new session, and execute below command with process id of your second session:

SELECT pg_blocking_pids(5936);

The result is:

pg_blocking_pids
--------------------
{6720}

PostgreSQL is a powerful, open source object-relational database system. It has more than 15 years of active development and a proven architecture that has earned it a strong reputation for reliability, data integrity, and correctness.
 
I am going to share few steps to install PostgreSQL 9.6 on Ubuntu 16.04.

First, check the version of Ubuntu:

lsb_release -sc

sudo add-apt-repository "deb https://apt.postgresql.org/pub/repos/apt/ trusty-pgdg main"

sudo apt-get update
sudo apt-get install postgresql-9.6

ubuntu@:~$ sudo passwd postgres
Enter new UNIX password:****
Retype new UNIX password:****
passwd: password updated successfully

sudo service postgresql start

ubuntu@:~$ su postgres
Password:****

createdb database_name

psql -d database_name

create table Test (rno integer);

\q

exit

A most awaited feature, The Powerful Parallel Sequential Query Scan feature introduced in PostgreSQL 9.6.

You can easily set the Parallel Sequential parameter’s value and can execute your queries 10 times faster.
When we are talking about Parallel Sequential Scanning, in background multiple workers or CPU threads are responsible for executing one single query.

In the PostgreSQL 9.6, You can change the Process Workers parameter value which is default 8. (Using max_worker_processes parameter)

Important Note:
If you want to increase Parallel Process Worker parameter value, You should make sure about your total CPU usage because More Parallel Process Worker required More CPU.
Once you change the parameter value, You must restart the PostgreSQL service.

Database Theory: What is Parallel Query Processing (Parallel Database System)?

The details about new Parallel Sequential Parameters:

max_worker_processes: Total number of background workers limited by this parameter. (Default: 8)
max_parallel_workers_per_gather: Total number of workers that can assist a sequential scan.
parallel_setup_cost: used to estimate the cost of instantiate a worker.
min_parallel_relation_size: You can define the minimum size of the relation and only those relations will consider for additional workers.
parallel_tuple_cost: used to estimate the cost of transferring a tuple from one worker to another.

Below is a full demonstration of this:

Prepared different result set by changing the max_parallel_workers_per_gather parameter value and once you change the value of this parameter, please do not forget to restart PostgreSQL Service. (You can find this parameter in postgresql.conf)

For each execution result, You should compare the value of “Execution time”.’

First create one sample table:

CREATE TABLE tbl_ItemTransactions 
 (
     TranID SERIAL
     ,TransactionDate TIMESTAMPTZ
     ,TransactionName TEXT
 );

INSERT INTO tbl_ItemTransactions 
(TransactionDate, TransactionName)
SELECT x, 'dbrnd' 
FROM generate_series('2014-01-01 00:00:00'::timestamptz, '2016-08-01 00:00:00'::timestamptz,'2 seconds'::interval) a(x);

SELECT pg_size_pretty(pg_total_relation_size('tbl_ItemTransactions')) AS TableSize;
-- total size "2028 MB"

EXPLAIN ANALYZE SELECT COUNT(1) FROM tbl_ItemTransactions;

"Aggregate  (cost=768697.65..768697.66 rows=1 width=8) 
(actual time=59863.028..59863.029 rows=1 loops=1)"
"  ->  Seq Scan on tbl_itemtransactions  
(cost=0.00..666853.32 rows=40737732 width=0) 
(actual time=0.825..30647.470 rows=40737601 loops=1)"
"Planning time: 0.079 ms"
"Execution time: 59863.095 ms"

EXPLAIN ANALYZE SELECT COUNT(1) FROM tbl_ItemTransactions;

"Finalize Aggregate  (cost=472651.90..472651.91 rows=1 width=8) 
(actual time=21129.508..21129.511 rows=1 loops=1)"
"  ->  Gather  (cost=472651.69..472651.90 rows=2 width=8) 
(actual time=21129.213..21129.492 rows=3 loops=1)"
"        Workers Planned: 2"
"        Workers Launched: 2"
"        ->  Partial Aggregate  (cost=471651.69..471651.70 rows=1 width=8) 
(actual time=21102.175..21102.176 rows=1 loops=3)"
"              ->  Parallel Seq Scan on tbl_itemtransactions  
(cost=0.00..429216.55 rows=16974055 width=0) 
(actual time=0.048..10929.919 rows=13579200 loops=3)"
"Planning time: 0.052 ms"
"Execution time: 21132.005 ms"

EXPLAIN ANALYZE SELECT COUNT(1) FROM tbl_ItemTransactions;

"Finalize Aggregate  (cost=387781.83..387781.84 rows=1 width=8) 
(actual time=16212.360..16212.361 rows=1 loops=1)"
"  ->  Gather  (cost=387781.41..387781.82 rows=4 width=8) 
(actual time=16211.487..16212.347 rows=5 loops=1)"
"        Workers Planned: 4"
"        Workers Launched: 4"
"        ->  Partial Aggregate  (cost=386781.41..386781.42 rows=1 width=8) 
(actual time=16159.409..16159.409 rows=1 loops=5)"
"              ->  Parallel Seq Scan on tbl_itemtransactions  
(cost=0.00..361320.33 rows=10184433 width=0) 
(actual time=0.043..9301.194 rows=8147520 loops=5)"
"Planning time: 0.050 ms"
"Execution time: 13516.982 ms"

EXPLAIN ANALYZE SELECT COUNT(1) FROM tbl_ItemTransactions;

"Finalize Aggregate  (cost=345346.90..345346.91 rows=1 width=8) 
(actual time=16514.376..16514.377 rows=1 loops=1)"
"  ->  Gather  (cost=345346.27..345346.88 rows=6 width=8) 
(actual time=16511.341..16514.351 rows=7 loops=1)"
"        Workers Planned: 6"
"        Workers Launched: 6"
"        ->  Partial Aggregate  (cost=344346.27..344346.28 rows=1 width=8) 
(actual time=16358.386..16358.387 rows=1 loops=7)"
"              ->  Parallel Seq Scan on tbl_itemtransactions  
(cost=0.00..327372.22 rows=6789622 width=0) 
(actual time=0.055..9956.567 rows=5819657 loops=7)"
"Planning time: 0.034 ms"
"Execution time: 10288.810 ms"