intro to postgresql - joeconway.com · intro to postgresql joe conway [email protected]...

Intro to PostgreSQL

Joe [email protected]

[email protected]

Crunchy Data

March 03, 2017

InstallationConfiguration

SQLAppendix

TypicalDebug Build From SourceCluster InitializationPostgreSQL Service Control

Binary Packages

RPMs

Debian packages

Windows installer

Source?

Joe Conway PostgreSQL@SCaLE15x 2/89


SQLAppendix


Build Dependencies

On Red Hat 7.x (or variant)

git (PostgreSQL source tree)Build Dependencies (might not be all inclusive)

yum install git bison flex gcc ccache readline-devel \

zlib-devel perl-devel perl-ExtUtils-Embed openssl-devel \

pam-devel libxml2-devel libxslt-devel libuuid-devel \

openldap-devel tcl-devel python-devel gdb



SQLAppendix


Clone Source

mkdir -p /opt/src/pgsql-git

cd /opt/src/pgsql-git/

git clone git://git.postgresql.org/git/postgresql.git

cd postgresql

# checkout the 9.5 tree

git checkout REL9_5_STABLE

http://www.postgresql.org/docs/9.5/static/git.html


http://www.postgresql.org/docs/9.5/static/git.html


SQLAppendix


Configure and Build

Already have PGDG RPMs installed

Want very similarly configured source tree

Capture and modify CFLAGS⇒ Change -O2 -g to -O0 -g3

⇒ Remove -Wp,-D\_FORTIFY\_SOURCE=2

export PATH=/usr/pgsql-9.5/bin:$PATH

eval ./configure $(pg_config --configure | \

sed -e 's/-O2 -g/-O0 -g3/g' \

-e 's/-Wp,-D\_FORTIFY\_SOURCE=2//g')

make -j 8

make install



SQLAppendix


Debug Workflow

Log in using psql

Find backend process pid

Attach with gdb

Set breakpoint

Execute problematic SQL statement



SQLAppendix


initdb

Source build/do-it-yourself<path-to-postgres>/bin/initdb -D $PGDATA

Red Hat/CentOS 7.x

/usr/pgsql-9.5/bin/postgresql95-setup initdb

creates cluster and config files in /var/lib/pgsql/9.5/data/

Debian-based

pg_createcluster 9.5 main

creates cluster in /var/lib/postgresql/9.5/main

config files created in /etc/postgresql/9.5/main



SQLAppendix


Starting PostgreSQL

Single user modepostgres --single -D $PGDATA dbname

Manualpg_ctl -D $PGDATA -l /path/to/logfile start

Red Hat/CentOS 6.xservice postgresql start

Red Hat/CentOS 7.xsystemctl start postgresql-9.5

Debian-basedpg_ctlcluster 9.5 main start



SQLAppendix


Stopping PostgreSQL

Single user modeControl+D (type EOF character)

Manualpg_ctl -D $PGDATA stop -m fast

Red Hat/CentOS 6.xservice postgresql stop

Red Hat/CentOS 7.xsystemctl stop postgresql-9.5

Debian-basedpg_ctlcluster 9.5 main stop



SQLAppendix


Shutdown Modes

smart: wait until existing sessions exit

fast: gracefully terminate existing sessions (default)

immediate: kill all processes

pg_ctl -D $PGDATA stop -m immediate



SQLAppendix


Terminate Particular SessionIn bash terminal:

ps -fu postgres |grep test

postgres 30999 1837 0 16:56 ? 00:00:00 postgres: postgres test [local] idle

kill -SIGTERM 30999

In psql:

SELECT pid, state, clock_timestamp() - state_change as age, query

FROM pg_stat_activity WHERE datname = 'test';

pid | state | age | query

-------+---------------------+-----------------+--------

31255 | idle in transaction | 00:00:26.020821 | begin;

(1 row)

SELECT pg_terminate_backend(31255);



SQLAppendix


Cancel Long Running QueriesIn bash terminal:

ps -fu postgres |grep test

postgres 30999 1837 0 16:56 ? 00:00:00 postgres: postgres test [local] idle

kill -SIGINT 30999

In psql:

SELECT pid, state, clock_timestamp() - state_change as age, query

FROM pg_stat_activity WHERE datname = 'test';

pid | state | age | query

-------+---------------------+-----------------+--------

31255 | idle in transaction | 00:00:26.020821 | begin;

(1 row)

SELECT pg_cancel_backend(31255);



SQLAppendix

postgresql.confpg hba.conf

Configuration

File postgresql.conf and postgresql.auto.conf:

# comment

name = value

To activate configuration changes:

psql -c "SELECT pg_conf_reload()" # as postgres

pg_ctl -D $PGDATA reload # as postgres

service postgresql reload # typically as root

systemctl reload postgresql-9.5 # typically as root

kill -SIGHUP <postmaster-pid> # as postgres or root

http://www.postgresql.org/docs/current/interactive/runtime-config.html


http://www.postgresql.org/docs/current/interactive/runtime-config.html


SQLAppendix


Configuration

Persistent

Modify: postgresql.conf

Add: postmaster options (e.g. in startup script)With SQL:

ALTER SYSTEM SET, ALTER SYSTEM RESET

ALTER [DATABASE] <dbname> SET <var> = <val>;

ALTER [ROLE] <rolename> SET <var> = <val>;

Per Session (SQL)

SET, RESET, SHOWSELECT * FROM pg_settings;

UPDATE pg_settings SET setting = '<val>' WHERE name = '<var>';

SELECT current_setting('<var>');

SELECT set_config('<var>','<val>','<local T/F>');



SQLAppendix


Connection Management

listen_addresses = 'localhost'

port = 5432

max_connections = 100

superuser_reserved_connections = 2

ssl = false

http://www.postgresql.org/docs/current/interactive/runtime-config-connection.html


http://www.postgresql.org/docs/current/interactive/runtime-config-connection.html


SQLAppendix


Memory Management

shared_buffers = 128MB

huge_pages = try

work_mem = 4MB

maintenance_work_mem = 64MB

shared_preload_libraries = ''

http://www.postgresql.org/docs/current/interactive/runtime-config-resource.html


http://www.postgresql.org/docs/current/interactive/runtime-config-resource.html


SQLAppendix


Write-Ahead Log

wal_level = minimal

fsync = on

synchronous_commit = on

checkpoint_timeout = 5min

# checkpoint_segments = 3 (pre-9.5)

max_wal_size = 1GB

min_wal_size = 80MB

checkpoint_completion_target = 0.5

checkpoint_warning = 30s

http://www.postgresql.org/docs/current/interactive/runtime-config-wal.html


http://www.postgresql.org/docs/current/interactive/runtime-config-wal.html


SQLAppendix


Planner

effective_cache_size = 128MB

seq_page_cost = 1.0

random_page_cost = 4.0

http://www.postgresql.org/docs/current/interactive/runtime-config-query.html


http://www.postgresql.org/docs/current/interactive/runtime-config-query.html


SQLAppendix


Logging

log_destination = 'stderr' # stderr, syslog, csvlog, eventlog

logging_collector = off

log_directory = 'pg_log'

log_filename = 'postgresql-%Y-%m-%d_%H%M%S.log'

log_file_mode = 0600

log_truncate_on_rotation = false

log_rotation_age = 1d

log_rotation_size = 10MB

syslog_facility = 'LOCALO'

syslog_ident = 'postgres'

http://www.postgresql.org/docs/current/interactive/runtime-config-logging.html


http://www.postgresql.org/docs/current/interactive/runtime-config-logging.html


SQLAppendix


Logging

log_min_duration_statement = -1

log_statement = 'none' # none, mod, ddl, all

log_connections = off

log_disconnections = off

log_hostname = off

log_timezone = ''

log_lock_waits = off

log_temp_files = -1

log_checkpoints = off

log_autovacuum_min_duration = -1



SQLAppendix


Logging

log_line_prefix = '' # special values:

# %a = application name, %u = user name,

# %d = database name, %r = remote host and port,

# %p = process ID, %m = timestamp with milliseconds,

# %i = command tag, %e = SQL state, %c = session ID,

# %l = session line number, ... and others



SQLAppendix


Miscellaneous

search_path = '$user,public'

statement_timeout = 0

datestyle = 'iso, mdy'

timezone = unknown

http://www.postgresql.org/docs/current/interactive/runtime-config-client.html


http://www.postgresql.org/docs/current/interactive/runtime-config-client.html


SQLAppendix


Host Based Authentication File

Which hosts are allowed to connect

How clients are authenticated

Which PostgreSQL user names they can use

Which databases they can access

https://www.postgresql.org/docs/9.5/static/auth-pg-hba-conf.html




SQLAppendix



Read on server startup

Must reload postmaster for changes to take effect

First line matching conn type, address, database, and user is used for authentication

If line picked and authentication fails, access denied

If no line matches, access denied





SQLAppendix



Lines (rules/records) look like this

# CONN-TYPE DATABASE USER ADDRESS METHOD OPTIONS

# local <dbname> <user> <method> [<opts>]

# host <dbname> <user> <address> <method> [<opts>]

# hostssl <dbname> <user> <address> <method> [<opts>]

# hostnossl <dbname> <user> <address> <method> [<opts>]



SQLAppendix


Connection Type

Specifies type of connection the rule matches

local: Unix-domain socket

host: Either plain or SSL-encrypted TCP/IP socket

hostssl: SSL-encrypted TCP/IP socket

hostnossl: Plain TCP/IP socket



SQLAppendix


Database

Specifies set of databases the rule matches

all: Wildcard

sameuser: Database name matches user name

samerole: User part of role/group matching database name

replication: all keyword does not match replication

<dbname>[,<dbname>]: One or more specific database names

@<filename>: Separate file containing names to match



SQLAppendix


User

Specifies set of users the rule matches

all: Wildcard

<username>[,<username>]: One or more user names

+<groupname>: Any roles that are directly or indirectly members of this role

@<filename>: Separate file containing names to match



SQLAppendix


Address

Specifies set of client hosts the rule matches

<IPAddr>/<CIDR-Mask>: Host or Network

<IPAddr> <Mask>: Host or Network

[.]<hostname>: [suffix] actual FQ hostname

samehost: match any of server’s own IP addresses

samenet: match any address in any subnet that server directly connected to



SQLAppendix


Method

Specifies authentication method to use when connection matches rule

trust

md5, password

cert

peer, pam, ident

gss, sspi, ldap, radius

reject

https://www.postgresql.org/docs/9.5/static/auth-methods.html


https://www.postgresql.org/docs/9.5/static/auth-methods.html


SQLAppendix


Options

Set of options for the authentication in the format NAME=VALUE

Options depend authentication method

Refer to ”Client Authentication” section of docs



SQLAppendix


Brief Example

local all all trust

host all all 127.0.0.1/32 trust

host all all samenet md5

host all all 192.168.1.0/24 ldap ldapurl="ldap://ldap.ex.net/dc=ex,dc=net?uid?sub"



SQLAppendix

General

psql - Hints

psql [-h <host> -p <port> -U <user>] <dbname> [-c "some sql"]

watch -n1 "psql <dbname> -c 'some sql'"

psql -l

psql -E

https://www.postgresql.org/docs/9.5/static/app-psql.html


https://www.postgresql.org/docs/9.5/static/app-psql.html


SQLAppendix

General

psql - Hints

<Up>

<Ctrl>-r

<tab>



SQLAppendix

General

psql - Hints

\?

\h

\dt [<schemaname>].[<tablename>][*]

\d <tablename>[*]

\df <funcname>[*]

\x

\e

\watch [ seconds ]



SQLAppendix

General

General Notes on Syntax

Identifier:

Unquoted: name (lower cased)Quoted: "Name" (preserved)Embedded Quotes: "List ""A"""

Unicode identifier: U&"\0441\043B\043E\043D"

String Literal:

Simple: 'text'

Dollar: see next slideEmbedded Quotes: 'McDonald''s restaurant'

Unicode literal: U&'\0441\043B\043E\043D'

Comments:

Single line: -- comment

Multi-line: /* comment */



SQLAppendix

General

Dollar Quoting

$<tag>$

<tag> is zero or more characters

Start and End tag must match

Particularly useful for function bodies

Works for all character strings

Nest by choosing different <tag> at each level

CREATE OR REPLACE FUNCTION dummy()

RETURNS text AS $_$

BEGIN

RETURN $$Say 'hello'$$;

END;

$_$ LANGUAGE plpgsql;

http://www.postgresql.org/docs/9.5/static/sql-syntax-lexical.html#SQL-SYNTAX-DOLLAR-QUOTING


http://www.postgresql.org/docs/9.5/static/sql-syntax-lexical.html#SQL-SYNTAX-DOLLAR-QUOTING


SQLAppendix

General

Data Types - Character Strings

text: character string with variable length without upper limit

varchar(X): character string, at most X characters long

char(X): character string, exactly X characters long (padded with spaces)



SQLAppendix

General

Data Types - Numbers

smallint: integer (2 bytes) (alias: int2)

integer: integer (4 bytes) (alias: int4, int)

bigint: integer (8 bytes) (alias: int8)

serial: not a real datatype – int with a default value expression that automatically takes thenext value from a sequence

bigserial: the same using bigint as base type

real: floating-point number (4 bytes) (alias: float4)

double precision: floating-point number (8 bytes) (alias: float8)

numeric(M,N): fixed-point number, max. M digits total, thereof N digits after the decimal point



SQLAppendix

General

Data Types - Date/Time

date: date (without time)

time [without time zone]: time without time zone

time with time zone: time with time zone

timestamp [without time zone]: date and time without time zone

timestamp with time zone: date and time with time zone

interval: interval length



SQLAppendix

General

Data Types - Enumeration Types

Real enumeration type

Seamless operator and function support

CREATE TYPE color AS enum('red', 'green', 'blue');

CREATE TABLE clothes (type text, color color);

INSERT INTO clothes VALUES('shirt', 'red');



SQLAppendix

General

Data Types - Geometric

Don’t use native, use PostGIS instead

Out of scope for this talk

CREATE EXTENSION postgis;



SQLAppendix

General

Data Types - JSON

Checks well-formedness

Compliant JSON requires UTF-8 server encoding

Stored an exact copy of the input text as a string

Preserves semantically-insignificant white space between tokensPreserves order of keys within JSON objectsIf contains same key more than once, all key/value pairs are kept



SQLAppendix

General

Data Types - JSONB


Compliant JSON requires UTF-8 server encoding

Full indexing

Stored in a decomposed binary format

Does not preserve white spaceDoes not preserve the order of object keysDoes not keep duplicate object keys (only the last value is kept)

https://www.postgresql.org/docs/9.5/static/datatype-json.html


https://www.postgresql.org/docs/9.5/static/datatype-json.html


SQLAppendix

General

Data Types - XML Type


Requires additional support from the underlying operating system (libxml2, libxslt)

Needs some care regarding encoding issues (UTF-8 vs. others)

No comparison or indexable operators available

No DTD validation



SQLAppendix

General

Data Types - Additional

bytea: binary data

boolean: true or false

array: multidimensional arrays from elements of same scalar data type

composite: tuples from elements of any data type

range: represent a range of values of some element type

text search: designed to support full text search

other: many other types for network addresses, uuid, . . .

https://www.postgresql.org/docs/9.5/static/datatype.html


https://www.postgresql.org/docs/9.5/static/datatype.html


SQLAppendix

General

Thank You

Questions?



SQLAppendix

Data DefinitionData ModificationQueries

CREATE TABLE

CREATE [ [ GLOBAL | LOCAL ] { TEMPORARY | TEMP } | UNLOGGED ] TABLE [ IF NOT EXISTS ]

table_name ( [

{ column_name data_type [ COLLATE collation ] [ column_constraint [ ... ] ]

| table_constraint

| LIKE source_table [ like_option ... ] }

[, ... ]

] )

[ INHERITS ( parent_table [, ... ] ) ]

[ WITH ( storage_parameter [= value] [, ... ] ) | WITH OIDS | WITHOUT OIDS ]

[ ON COMMIT { PRESERVE ROWS | DELETE ROWS | DROP } ]

[ TABLESPACE tablespace_name ]

https://www.postgresql.org/docs/9.5/static/sql-createtable.html


https://www.postgresql.org/docs/9.5/static/sql-createtable.html


SQLAppendix


CREATE TABLE - Simple

Every table should have a primary key

Primary key columns are not-null implicitly

serial: auto-incrementing pseudotype

bigserial: auto-incrementing 64-bit integer pseudotype

CREATE TABLE category (

id serial PRIMARY KEY,

genre text NOT NULL,

description text

);



SQLAppendix


CREATE TABLE - Table Constraint

Every column constraint can be written as a table constraint

But not vice versa; example later

CREATE TABLE publisher (

id serial,

name text NOT NULL,

country text NOT NULL,

comment text,

PRIMARY KEY (id)

);



SQLAppendix


CREATE TABLE - Check Constraint

CREATE TABLE author (


lastname text NOT NULL,

firstname text NOT NULL,

primary_email text

CHECK (primary_email ~* '^[A-Za-z0-9._%-]+@[A-Za-z0-9.-]+[.][A-Za-z]+$'),

comment text

);



SQLAppendix


CREATE TABLE - Foreign Key

CREATE TABLE book (


title text NOT NULL,

subtitle text,

edition text NOT NULL,

isbn text NOT NULL,

description text,

language text,

publisher_id integer NOT NULL REFERENCES publisher (id)

);



SQLAppendix


CREATE TABLE - More Complex

CREATE TABLE bookauthor (

id serial,

book_id integer NOT NULL,

author_id integer NOT NULL,

comment text,

PRIMARY KEY (book_id, author_id),

FOREIGN KEY (book_id) REFERENCES book (id),

FOREIGN KEY (author_id) REFERENCES author (id)

);



SQLAppendix


DROP TABLE

DROP TABLE [ IF EXISTS ] name [, ...] [ CASCADE | RESTRICT ]

IF EXISTS: Do not throw error if table does not exist

CASCADE: Automatically drop dependent objects

RESTRICT: Refuse to drop if objects depend on it (default)



SQLAppendix


ALTER TABLE

ALTER TABLE [ IF EXISTS ] [ ONLY ] name [ * ]

action [, ... ]


RENAME [ COLUMN ] column_name TO new_column_name


RENAME CONSTRAINT constraint_name TO new_constraint_name

ALTER TABLE [ IF EXISTS ] name

RENAME TO new_name

ALTER TABLE [ IF EXISTS ] name

SET SCHEMA new_schema

ALTER TABLE ALL IN TABLESPACE name [ OWNED BY role_name [, ... ] ]

SET TABLESPACE new_tablespace [ NOWAIT ]

Changes definition of existing table

Several subforms, some examples follow

https://www.postgresql.org/docs/9.5/static/sql-altertable.html


https://www.postgresql.org/docs/9.5/static/sql-altertable.html


SQLAppendix


ALTER TABLE ADD COLUMN

ALTER TABLE book

ADD COLUMN publisher_id integer NOT NULL;

ALTER TABLE book

ADD COLUMN price numeric(7, 2)

CONSTRAINT pos_price CHECK (price > 0);

New column is always added as last in column list

Changing logical or physical order not supported (yet)



SQLAppendix


ALTER TABLE DROP COLUMN

ALTER TABLE book DROP COLUMN author_id;

ALTER TABLE book DROP COLUMN author_id CASCADE;



SQLAppendix


ALTER TABLE - Constraints

ALTER TABLE publisher ADD PRIMARY KEY (id);

ALTER TABLE book

ADD FOREIGN KEY (publisher_id) REFERENCES publisher (id);

ALTER TABLE publisher

ADD CONSTRAINT name_unique UNIQUE (name);

ALTER TABLE publisher

ADD CONSTRAINT check_postcode

CHECK (postcode SIMILAR TO '[0-9][0-9][0-9][0-9][0-9]');

Constraint names are optional but useful for referencing constraints⇒ e.g. to DROP them



SQLAppendix


ALTER TABLE - Changing Not-Null

ALTER TABLE book ALTER COLUMN edition SET NOT NULL;

ALTER TABLE book ALTER COLUMN edition DROP NOT NULL;



SQLAppendix


ALTER TABLE - Changing Column Type

ALTER TABLE book ALTER COLUMN price

TYPE numeric(7,2);

ALTER TABLE ALTER COLUMN number

TYPE text USING to_char(number, 'RN');

Implicit type casting possible for certain data types only

Otherwise provide explicit type casting expression



SQLAppendix


ALTER TABLE RENAME

ALTER TABLE books RENAME TO ebooks;



SQLAppendix


ALTER TABLE RENAME COLUMN

ALTER TABLE number_representation

RENAME COLUMN number TO roman_numerals;



SQLAppendix


INSERT

[ WITH [ RECURSIVE ] with_query [, ...] ]

INSERT INTO table_name [ AS alias ] [ ( column_name [, ...] ) ]

{ DEFAULT VALUES |

VALUES ( { expression | DEFAULT } [, ...] ) [, ...] | query }

[ ON CONFLICT [ conflict_target ] conflict_action ]

[ RETURNING * | output_expression [ [ AS ] output_name ] [, ...] ]

https://www.postgresql.org/docs/current/static/sql-insert.html


https://www.postgresql.org/docs/current/static/sql-insert.html


SQLAppendix


UPDATE


UPDATE [ ONLY ] table_name [ * ] [ [ AS ] alias ]

SET { column_name = { expression | DEFAULT } |

( column_name [, ...] ) = ( { expression | DEFAULT } [, ...] ) |

( column_name [, ...] ) = ( sub-SELECT )

} [, ...]

[ FROM from_list ]

[ WHERE condition | WHERE CURRENT OF cursor_name ]


https://www.postgresql.org/docs/current/static/sql-update.html


https://www.postgresql.org/docs/current/static/sql-update.html


SQLAppendix


UPDATE

UPDATE book SET price = 42.24

WHERE lower(title) = 'postgresql';

UPDATE author SET lastname='Bar', firstname='Foo'

WHERE lastname='Doe';



SQLAppendix


DELETE


DELETE FROM [ ONLY ] table_name [ * ] [ [ AS ] alias ]

[ USING using_list ]

[ WHERE condition | WHERE CURRENT OF cursor_name ]


https://www.postgresql.org/docs/current/static/sql-delete.html


https://www.postgresql.org/docs/current/static/sql-delete.html


SQLAppendix


DELETE

DELETE FROM publisher WHERE name LIKE 'Publisher %';

DELETE FROM publisher WHERE id > 1;

DELETE FROM author WHERE primary_email ~ '.*@example.com';

http://www.postgresql.org/docs/9.3/interactive/sql-delete.html


http://www.postgresql.org/docs/9.3/interactive/sql-delete.html


SQLAppendix


Queries


SELECT [ ALL | DISTINCT [ ON ( expression [, ...] ) ] ]

[ * | expression [ [ AS ] output_name ] [, ...] ]

[ FROM from_item [, ...] ] [ WHERE condition ]

[ GROUP BY grouping_element [, ...] ] [ HAVING condition [, ...] ]

[ WINDOW window_name AS ( window_definition ) [, ...] ]

[ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] select ]

[ ORDER BY expression [ ASC | DESC | USING operator ]

[ NULLS { FIRST | LAST } ] [, ...] ]

[ LIMIT { count | ALL } ] [ OFFSET start [ ROW | ROWS ] ]

[ FETCH { FIRST | NEXT } [ count ] { ROW | ROWS } ONLY ]

[ FOR { UPDATE | NO KEY UPDATE | SHARE | KEY SHARE }

[ OF table_name [, ...] ] [ NOWAIT | SKIP LOCKED ] [...] ]

https://www.postgresql.org/docs/current/static/sql-select.html


https://www.postgresql.org/docs/current/static/sql-select.html


SQLAppendix


Queries

SELECT * FROM author;

SELECT name AS "Publisher" FROM publisher;

SELECT DISTINCT author_id FROM bookauthor;

SELECT title FROM book WHERE price > 10.00;



SQLAppendix


Queries

SELECT * FROM book

WHERE lower(title) LIKE '%sql%'

ORDER BY price DESC;

SELECT b.title

FROM bookauthor ba, author a, book b

WHERE a.lastname = 'Bar'

AND ba.author_id = a.id

AND b.id = ba.book_id

ORDER BY 1 DESC;



SQLAppendix


Joins

CROSS JOIN

INNER JOIN

OUTER JOIN

LEFT

RIGHT

FULL



SQLAppendix


CROSS JOIN

Join each row of t1 with each row of t2

Usually written in error (first form below)

Occasionally useful (use second form to make intention clear)

SELECT * FROM t1, t2;

SELECT * FROM t1 CROSS JOIN t2;



SQLAppendix


INNER JOIN

Join each row of t1 with each row of t2 which matches condition

condition typically something like t1.cx = t2.cy

USING implies match on same-named columns from t1 and t2

USING also means only one of each pair included in join output

NATURAL same as USING that mentions all columns in both tables with same names

SELECT ... FROM t1, t2 WHERE condition;

SELECT ... FROM t1 [ INNER ] JOIN t2 ON condition;

SELECT ... FROM t1 [ INNER ] JOIN t2 USING (column list);

SELECT ... FROM t1 NATURAL [ INNER ] JOIN t2;



SQLAppendix


INNER JOIN

SELECT * FROM book AS b, publisher AS p

WHERE b.publisher_id = p.id

AND b.title ilike '%post%';

SELECT * FROM book AS b JOIN publisher AS p

ON b.publisher_id = p.id

WHERE b.title ilike '%post%';



SQLAppendix


OUTER JOIN

LEFT JOIN: all rows from the left table

RIGHT JOIN: all rows from the right table

FULL JOIN: all rows from both tables

Join each row of t1 with each row of t2 which matches condition

Non-matching rows on ”all” side added to the result⇒ filled out with NULLs for missing side

SELECT ... FROM t1 LEFT/RIGHT/FULL [ OUTER ] JOIN t2 ON condition;

SELECT ... FROM t1 LEFT/RIGHT/FULL [ OUTER ] JOIN t2 USING (column list);

SELECT ... FROM t1 NATURAL LEFT/RIGHT/FULL [ OUTER ] JOIN t2;



SQLAppendix


OUTER JOIN

SELECT * FROM publisher LEFT JOIN book

ON publisher.id = book.publisher_id;



SQLAppendix


Subqueries - Uncorrelated

Calculates constant result set for outer query

Executed once

SELECT

author.lastname

,author.firstname

FROM

bookauthor

,book

,author

WHERE

book.id IN (SELECT id FROM book WHERE title LIKE '%Spam%')

AND bookauthor.book_id = book.id

AND author.id = bookauthor.author_id;



SQLAppendix


Subqueries - Correlated

References variables from outer query

Repeated for each row of outer query

Could be rewritten as JOIN

SELECT

book_id

,author_id

FROM

bookauthor ba

WHERE

book_id = (SELECT id FROM book

WHERE id = ba.book_id);



SQLAppendix


CREATE VIEW

Views are just saved SQL queries

No precalculation, no special optimizations

Provide notational convenience

Simple views are automatically updatable - INSERT/UPDATE/DELETE allowed if:

Has exactly one entry in FROM list, which must be table or another updatable viewMust not contain WITH, DISTINCT, GROUP BY, HAVING, LIMIT, or OFFSETclauses at the top levelMust not contain set operations (UNION, INTERSECT or EXCEPT) at the toplevelMust not contain any aggregates, window functions or set-returning functions



SQLAppendix


CREATE VIEW

CREATE [ OR REPLACE ] [ TEMP | TEMPORARY ] [ RECURSIVE ]

VIEW name [ ( column_name [, ...] ) ]

[ WITH ( view_option_name [= view_option_value] [, ... ] ) ]

AS query

[ WITH [ CASCADED | LOCAL ] CHECK OPTION ]



SQLAppendix


CREATE VIEW

CREATE VIEW booklist AS

SELECT book.title, author.lastname, author.firstname

FROM bookauthor, book, author

WHERE bookauthor.book_id = book.id




SQLAppendix


DROP VIEW

DROP VIEW [ IF EXISTS ] name [, ...] [ CASCADE | RESTRICT ]



SQLAppendix


CREATE TABLE AS

Creates a copy

Data changes are specific to the new table

CREATE TABLE booklist AS

SELECT book.title, author.lastname, author.firstname

FROM bookauthor, book, author

WHERE bookauthor.book_id = book.id




SQLAppendix


Common Table Expressions

Expressions declared by WITH

Also known as ”CTE” or ”WITH clause”

Can contain recursive references to itself

Can be thought of as in-memory temp table, private to query



SQLAppendix


CTE - Syntax


Where with_query is:

with_query_name [ ( column_name [, ...] ) ]

AS ( select | values | insert | update | delete )

When RECURSIVE specified subquery must have the form:

non_recursive_term UNION [ ALL | DISTINCT ] recursive_term

https://www.postgresql.org/docs/current/static/queries-with.html


https://www.postgresql.org/docs/current/static/queries-with.html


SQLAppendix


CTE - Simple non-recursive example

WITH t(a) AS

(

SELECT * FROM generate_series(1,10)

)

SELECT t.a * t.a FROM t;



SQLAppendix


CTE - Simple recursive example

WITH RECURSIVE t(n) AS

(

VALUES (1)

UNION ALL

SELECT n+1 FROM t WHERE n < 100

)

SELECT sum(n) FROM t;



SQLAppendix


CTE - DML example

CREATE TABLE t1(t1_id serial PRIMARY KEY, t1_f2 text);

CREATE TABLE t2(t2_id serial, t1_id int REFERENCES t1, t2_f3 text);

WITH

wt1(t1_id, t1_f2) AS

(INSERT INTO t1 VALUES(DEFAULT, 'some f2 text') RETURNING *),

wt2(t2_id, t1_id, t2_f3) AS

(INSERT INTO t2(t1_id, t2_f3) SELECT wt1.t1_id, 'some f3 text' FROM wt1 RETURNING *)

SELECT * FROM wt1 NATURAL JOIN wt2;

t1_id | t1_f2 | t2_id | t2_f3

-------+--------------+-------+--------------

1 | some f2 text | 1 | some f3 text

(1 row)


intro to postgresql - joeconway.com · intro to postgresql joe conway [email protected]...

Documents