This document provides an overview and agenda for a SQLAlchemy primer presentation. The presentation will cover myths about SQLAlchemy, core concepts like the engine and ORM, and include hands-on examples using the SQLite sakila database. The speaker is Yasushi Masuda, a Pythonista since 2001 and founder of PyCon JP, who will introduce SQLAlchemy basics and dispel common misconceptions.

1. SQLAlchemy Primer
(extra content)
for Kobe Python Meetup #13
2017/09/15 Kobe Japan

2. Yasushi Masuda PhD
( @whosaysni )
Tech team, Core IT grp. IT Dept.
MonotaRO Co., LTD.
Pythonista since 2001 (2.0～)
• elaphe (barcode library)
• oikami.py （老神.py）
• PyCon JP founder
Japanese Translation works

3. Agenda
Myths
Core concepts in SQLAlchemy
Engine basics (+hands-on)
ORM primer (+hans-on)

4. References
Online Document:
http://docs.sqlalchemy.org/
en/rel_1_1/
(Old) Japanese translation:
http://omake.accense.com/
static/doc-ja/sqlalchemy/

5. Preparation
sakila DB on SQLite http://bit.ly/2fdeeft
https://github.com/jOOQ/jOOQ/jOOQ-examples/Sakila/sqlite-sakila-db/sqlite-sakila.sq
Sakila
• Demonstration
DB for MySQL
• Models a rental
video shop
• BSD License
Schema described at:
https://dev.mysql.com/doc/sakila/en/sakila-structure-tables.html

6. Myths
[WRONG!] It's just an ORM library
Not limited to. SQLAlchemy is a DB manipulation framework.
[WRONG!] SA is built on ORM
NO. Connection management and SQL expression framework are its core.
ORM is built on them.
[WRONG!] SA cannot handle raw SQL
Table definition is not required. Even textual SQL is available.
[WRONG!] SA automatically escapes value for you
SQLAlchemy relies value escaping on DB-API, while it escapes schema, table, column
names. SQLAlchemy generates parameterized query that helps DB-API level escaping.
[WRONG!] Only Python adept can handle SA
SQLAlchemy is EASY. You may need SQL and RDBMS experience. Let's see!
[AGREE] Documentation is somewhat difficult to understand
Agreed.

7. Core concepts

8. http://docs.sqlalchemy.org/en/latest/index.html

9. http://docs.sqlalchemy.org/en/latest/index.html
SQLAlchemy Core
Engine (connection)
Schema definition
SQL Expression
SQLAlchemy ORM
Mapper
Declarative Mapper
Session
Dialect
DB Backend-specific functionalities

10. Engine
manages DB connection(s)
SQL Expression
describes SQL statement in Python
Mapping
reflects DB record with Python object
Dialect
DB Backend specific functionalities

11. Database
Your program
SQL
construction
Query
execution
Type
conversion
Parameter
binding
Driver
Setup
Connection
management
Result
data structure
Value
escaping
Schema
name
Type conversion
Dialect-
specific
Query Construction
Query Execution
Result
Schema object
Object-relational
mapping
High-level Interface
DB Session

12. Engine

13. DB API DB ServerProgram
Issues around DB-API
Which DB-API to use
How to initialize the DB-API
How to generate valid query for it
How to handle cursor on the DB-API
How to execute query on the DB-API
How to retrieve result from DB-API
How to reserve/reuse connection

14. DB API DB ServerProgram
Engine resolves issues
Select DB-API from DSN URL
Initialize DB-API for you
Accept string and SQL expression
Manage cursor for you
Unify execution/transaction API
Handle result via ResultProxy
Pool connection automatically
Engine

15. Engine DB API DB ServerProgram
>>> from sqlalchemy import create_engine
>>>

16. Engine DB API DB ServerProgram
>>> from sqlalchemy import create_engine
>>> e = create_engine('sqlite://') # SQLite memory engine
>>>

17. Engine DB API DB ServerProgram
# Use URL for specifying database
>>> e = create_engine('sqlite://') # SQLite memory engine
>>> e = create_engine('sqlite:///path_to_db.sqlite3') # sqlite3
>>> e = create_engine('mysql://scott:tiger@dbserv/dbname') # mysql
>>> e = create_engine('mssql://bill:gates@dbserv/dbname') # mssql
>>>

18. Engine DB API DB ServerProgram
# Be careful with number of slashes
>>> e = create_engine('sqlite:///sqlite-sakila.sq')
>>> e = create_engine('sqlite:////<absolute_path>/sqlite-sakila.sq')

19. Engine DB API DB ServerProgram
>>> e = create_engine('sqlite:///sqlite-sakila.sq')
>>> e
Engine(sqlite:///sakila-data.sq)
# execute returns ResultProxy
>>> q = 'select title from film limit 5')
>>> res = e.execute(q)
>>> res
<sqlalchemy.engine.result.ResultProxy object at 0x10da96990>
>>>

20. Engine DB API DB ServerProgram
>>> e = create_engine('sqlite:///sqlite-sakila.sq')
>>> e
Engine(sqlite:///sakila-data.sq)
# execute returns ResultProxy
>>> q = 'select title from film limit 5')
>>> res = e.execute(q)
>>> res
<sqlalchemy.engine.result.ResultProxy object at 0x10da96990
>>> for row in res: # ResultProxy can be iterated/namedtuple access
... print(row.title)
...
ACADEMY DINOSAUR
ACE GOLDFINGER
ADAPTATION HOLES
AFFAIR PREJUDICE
AFRICAN EGG
>>>

21. Engine DB API DB ServerProgram
>>> q = 'select film_id, title from film limit 10'
>>> rows = list(e.execute(q))
>>> rows[2][1] # each row is accessible like as tuple
'ADAPTATION HOLES'
>>> rows[4]['title'] # row can be accessible like as dictionary
'AFRICAN EGG'
>>> res = e.execute(q)
>>> for fid, title in res: # can be expanded as normal tuple
... print((fid, title))
...
(1, 'ACADEMY DINOSAUR')
(2, 'ACE GOLDFINGER')
(3, 'ADAPTATION HOLES')
(4, 'AFFAIR PREJUDICE')
(5, 'AFRICAN EGG')
>>> rows = list(e.execute(q))
>>>

22. transaction
>>> t = e.begin()
>>> t.transaction
<sqlalchemy...RootTransaction object at ...>
>>> t.transaction.commit()
# with statement handles transaction smart
>>> with e.begin():
e.execute(...)
>>> # (transaction committed automatically)

23. HANDS ON: Engine basics
Connect to sqlite-
sakila.sq database
List actors in film
"DINOSAUR SECRETARY"

24. HANDS ON: Engine basics
>>> e = create_engine('sqlite:///sqlite-sakila.sq')
>>> q = '''select a.first_name, a.last_name
... from film as f
... inner join film_actor as fa
... on f.film_id=fa.film_id
... inner join actor as a
... on fa.actor_id=a.actor_id
... where f.title = "DINOSAUR SECRETARY"'''
>>> for first_name, last_name in e.execute(q):
... print('{} {}'.format(first_name, last_name))
...
LUCILLE TRACY
BURT DUKAKIS
JAYNE NEESON
RUSSELL BACALL
PENELOPE MONROE
MINNIE KILMER

25. SQL expression

26. Remember: SQL is a language
SELECT [ ALL | DISTINCT [ ON ( expression [, ...] ) ] ]
[ * | expression [ [ AS ] output_name ] [, ...] ]
[ FROM from_item [, ...] ]
[ WHERE condition ]
[ GROUP BY grouping_element [, ...] ]
[ ORDER BY expression
[ ASC | DESC | USING operator ]
[ NULLS { FIRST | LAST } ] [, ...] ]
[ LIMIT { count | ALL } ]

27. Remember: SQL is a language
SELECT [ ALL | DISTINCT [ ON ( expression [, ...] ) ] ]
[ * | expression [ [ AS ] output_name ] [, ...] ]
[ FROM from_item [, ...] ]
[ WHERE condition ]
[ GROUP BY grouping_element [, ...] ]
[ ORDER BY expression
[ ASC | DESC | USING operator ]
[ NULLS { FIRST | LAST } ] [, ...] ]
[ LIMIT { count | ALL } ]
STATEMENT

28. Remember: SQL is a language
SELECT [ ALL | DISTINCT [ ON ( expression [, ...] ) ] ]
[ * | expression [ [ AS ] output_name ] [, ...] ]
[ FROM from_item [, ...] ]
[ WHERE condition ]
[ GROUP BY grouping_element [, ...] ]
[ ORDER BY expression
[ ASC | DESC | USING operator ]
[ NULLS { FIRST | LAST } ] [, ...] ]
[ LIMIT { count | ALL } ]
CLAUSE
CLAUSE
CLAUSE
CLAUSE
CLAUSE
CLAUSE

29. Remember: SQL is a language
SELECT [ ALL | DISTINCT [ ON ( expression [, ...] ) ] ]
[ * | expression [ [ AS ] output_name ] [, ...] ]
[ FROM from_item [, ...] ]
[ WHERE condition ]
[ GROUP BY grouping_element [, ...] ]
[ ORDER BY expression
[ ASC | DESC | USING operator ]
[ NULLS { FIRST | LAST } ] [, ...] ]
[ LIMIT { count | ALL } ]
parameter
parameter
expression
expression
expression
expression
expression
expression

30. Remember: SQL is a Language
SELECT
FROM selectables
selectable
join
WHERE
selectable
selectable
GROUP BY
ORDER BY
expression
expression
condition
expression
expressions
expression
expression
SELECT statement

31. SELECT A.name, B.title, C.title
FROM artists as A
INNER JOIN album as B ON A.id = B.artist_id
INNER JOIN music as C ON B.id = C.album
WHERE A.name LIKE '%Astor%'
AND A.year BETWEEN 1970 AND 2010
AND C.title LIKE '%Tango%'
From SQL to Python

32. SELECT <column>, <column>, ...
FROM <selectable>
INNER JOIN <selectable> ON <condition>
INNER JOIN <selectable> ON <condition>
...
WHERE <condition>
AND <condition>
AND <condition>
GROUP BY ...
LIMIT ...
Clauses

33. SELECT <[<column>, <column>, <column>]>
FROM <join (<selectable>, <selectable>, ...)>
WHERE <and (<condition>, <condition>, ...)>
Statement and subjects

34. SELECT <expressions>
FROM <selectable>
WHERE <conditions>
... simplifed

35. <select statement>
..., finally

36. engine.execute(<select statement>)
If query is "an object"...
Query object

37. >>> query = select(...)
>>> engine.execute(query)
... it can be "execute()-able"
Engine "compiles" query
into string and execute it
(according to dialect)

38. query = select(
<expression>,
from_obj=<selectables>,
whereclause=<conditions>,
)
clauses as parameters

39. columns = [col1, col2, ...]
fromobj = join(tbl1, tbl2, ...)
where = and_(expr1, expr2, ...)
query_expr = select(
columns, from_obj=fromobj,
whereclause=where)
query with SQL expression

40. >>> from sqlalchemy.sql import select, text
>>> q = select([text('*')])
building sql statement with
basic sql expression

41. >>> from sqlalchemy.sql import select, text
>>> q = select([text('*')])
>>> q
<sqlalchemy.....Select at ...; Select object>
>>> str(q)
'SELECT *'
>>> q = select([text('*')],
... from_obj=text('foo'),
... whereclause=text('id=3'))
>>> str(q)
'SELECT * nFROM foo nWHERE id=3'
building sql statement with
basic sql expression

42. >>> from sqlalchemy.sql import select, text
>>> q = select([text('*')])
>>> q
<sqlalchemy.....Select at ...; Select object>
>>> str(q)
'SELECT *'
>>> q = select([text('*')],
... from_obj=text('foo'),
... whereclause=text('id=3'))
>>> str(q)
'SELECT * nFROM foo nWHERE id=3'
building sql statement with
basic sql expression

43. elements: table and column
Schema
Table Table Table
...
Column
Column
Column
...
Column
Column
Column
...
Column
Column
Column
...
Schema
Table Table Table
Column
Column
Column
Column
Column
Column
...

44. elements: table and column
>>> from sqlalchemy.sql import column, table
>>> from sqlalchemy import INTEGER

45. elements: table and column
>>> from sqlalchemy.sql import column, table
>>> from sqlalchemy import INTEGER
>>> c = column('name') # simplest
>>> c = column('name', type_=INTEGER)

46. elements: table and column
>>> from sqlalchemy.sql import column, table
>>> from sqlalchemy import INTEGER
>>> c = column('name') # simplest
>>> c = column('name', type_=INTEGER)
>>> c
<sqlalchemy....ColumnClause at ...; name>
>>> str(c)
'name'

47. elements: table and column
>>> from sqlalchemy.sql import column, table
>>> from sqlalchemy import INTEGER
>>> c = column('name') # simplest
>>> c = column('name', type_=INTEGER)
>>> c
<sqlalchemy....ColumnClause at ...; name>
>>> str(c)
'name'
>>> c.table # None
>>> t1 = table('artist')
>>> c.table = t1
>>> str(c)
'artist.name'

48. >>> t = table('tbl1', column('col1'), ...)
defining table with columns
Table name List of columns

49. >>> t = table('tbl1', column('col1'), ...)
>>> t.c.col1
<sqlalchemy.....ColumnClause at ...; col1>
defining table with columns
Table name List of columns

50. >>> t = table('tbl1', column('col1'), ...)
>>> t.c.col1
<sqlalchemy.....ColumnClause at ...; col1>
>>> t.schema = 'db1'
>>> str(t.c.col1)
'db1.tbl1.col1'
defining table with columns
Table name List of columns

51. >>> t = table('tbl1', column('col1'), column('col2'))
select() with table element

52. >>> t = table('tbl1', column('col1'), column('col2'))
>>> print(select([t]))
SELECT tbl1.col1, tbl1.col2
FROM tbl1
select() with table element

53. >>> t = table('tbl1', column('col1'), column('col2'))
>>> print(select([t]))
SELECT tbl1.col1, tbl1.col2
FROM tbl1
# column labeling
>>> print(select([t.c.col1.label('col_alias1')])
SELECT tbl1.col1 as "col_alias1"
FROM tbl1
select() with table element

54. >>> t = table('tbl1', column('col1'), column('col2'))
>>> print(select([t]))
SELECT tbl1.col1, tbl1.col2
FROM tbl1
# column labeling
>>> print(select([t.c.col1.label('col_alias1')])
SELECT tbl1.col1 as "col_alias1"
FROM tbl1
# table alias
>>> t_A, t_B = alias(t, 'A'), alias(t, 'B')
>>> print(select([t_A.c.col1, t_B.c.col2]))
SELECT "A".col1, "B".col2
FROM tbl1 AS "A", tbl1 AS "B"
select() with table element

55. how to work with
where tbl1.col1 = 42
conditionals

56. how to work with
where tbl1.col1 = 42
conditionals
conditional expression
(compare operation)

57. >>> cond = text('last_name LIKE %KOV')
>>> str(cond)
'last_name LIKE %KOV'
conditional by text()

58. >>> cond = column('last_name').like('%KOV')
>>> cond
<sqlalchemy....BinaryExpression object at ...>
conditional by like() method

59. >>> cond = column('last_name').like('%KOV')
>>> cond
<sqlalchemy....BinaryExpression object at ...>
>>> str(cond)
'last_name LIKE :last_name_1'
conditional by like() method
placeholder for
right value of LIKE

60. >>> cond = column('last_name').like('%KOV')
>>> cond
<sqlalchemy....BinaryExpression object at ...>
>>> str(cond)
'last_name LIKE :last_name_1
>>> cond.right
BindParameter('%(4339977744 last_name)s',
'%KOV', type_=String())
conditional by like() method

61. >>> column('first_name') = 'DAVID'
File "<stdin>", line 1
SyntaxError: can't assign to function call
conditional by operation

62. >>> column('first_name') = 'DAVID'
File "<stdin>", line 1
SyntaxError: can't assign to function call
>>> column('first_name') == 'DAVID'
<sqlalchemy...BinaryExpression object at ...>
conditional by operation

63. >>> column('first_name') = 'DAVID'
File "<stdin>", line 1
SyntaxError: can't assign to function call
>>> column('first_name') == 'DAVID'
<sqlalchemy...BinaryExpression object at ...>
>>> cond = column('first_name') == 'DAVID'
>>> str(cond)
>>> 'first_name = :first_name_1'
>>> cond.right
>>> BindParameter('%(...)s', 'DAVID', type_=...)
conditional by operation

64. >>> column('first_name') = 'DAVID'
File "<stdin>", line 1
SyntaxError: can't assign to function call
>>> column('first_name') == 'DAVID'
<sqlalchemy...BinaryExpression object at ...>
>>> cond = column('first_name') == 'DAVID'
>>> str(cond)
>>> 'first_name = :first_name_1'
>>> cond.right
>>> BindParameter('%(...)s', 'DAVID', type_=...)
>>> str(column('last_name') == None)
>>> 'last_name is NULL'
conditional by operation

65. >>> from sqlalchemy.sql import select, table, column
>>> actor_tbl = table('actor', column('actor_id'),
... column('first_name'), column('last_name'))
conditionals in select

66. >>> from sqlalchemy.sql import select, table, column
>>> actor_tbl = table('actor', column('actor_id'),
... column('first_name'), column('last_name'))
>>> query = select([actor_tbl],
... whereclause=(actor_tbl.c.last_name=='TRACY'))
conditionals in select

67. >>> from sqlalchemy.sql import select, table, column
>>> actor_tbl = table('actor', column('actor_id'),
... column('first_name'), column('last_name'))
>>> query = select([actor_tbl],
... whereclause=(actor_tbl.c.last_name=='TRACY'))
>>> query
<sqlalchemy.....Select at ...; Select object>
>>> print(query)
SELECT actor.first_name, actor.last_name
FROM actor
WHERE actor.last_name = :last_name_1
>>> query.compile().params
{'last_name_1': 'TRACY'}
conditionals in select

68. >>> from sqlalchemy.sql import select, table, column
>>> actor_tbl = table('actor', column('actor_id'),
... column('first_name'), column('last_name'))
>>> query = select([actor_tbl],
... whereclause=(actor_tbl.c.last_name=='TRACY'))
>>> query
<sqlalchemy.....Select at ...; Select object>
>>> print(query)
SELECT actor.actor_id, actor.first_name, actor.last_name
FROM actor
WHERE actor.last_name = :last_name_1
>>> query.compile().params
{'last_name_1': 'TRACY'}
>>> list(e.execute(query))
[(20, 'LUCILLE', 'TRACY'), (117, 'RENEE', 'TRACY')]
conditionals in select

69. select(...)
SELECT ...
select(...).select_from(...)
SELECT ... FROM ...
select(...).where(...)
SELECT ... WHERE ...
select(...).where(...).where(...)
SELECT ... WHERE ... AND ...
select(...).where(...).order_by(...)
SELECT ... WHERE ... ORDER BY ...
generative method

70. actor_tbl.select()
SELECT ... FROM actor
actor_tbl.select(actor_tbl.c.first_name='BEN')
SELECT ... FROM actor WHERE first_name=...
generative method

71. HANDS ON: SQL expression
• define actor table with
table()/column()
• build query with sql
expression to search
actor having initial A.H.
(result should include
actor_id)

72. Schema definition

73. how to CREATE table
schema definition

74. how to CREATE table
how to define column detail
how to define constraints
schema definition

75. >>> from sqlalchemy import Column, MetaData, Table
>>> user_table = Table(
... 'user', MetaData(),
... Column('id', INTEGER, primary_key=True),
... Column('first_name', VARCHAR(45)),
... Column('last_name', VARCHAR(45)))
>>>
defining table

76. >>> from sqlalchemy import Column, MetaData, Table
>>> user_table = Table(
... 'user', MetaData(),
... Column('id', INTEGER, primary_key=True),
... Column('first_name', VARCHAR(45)),
... Column('last_name', VARCHAR(45)))
>>> user_table
Table('user', MetaData(bind=None), Column...)
defining table

77. >>> from sqlalchemy import Column, MetaData, Table
>>> user_table = Table(
... 'user', MetaData(),
... Column('id', INTEGER, primary_key=True),
... Column('first_name', VARCHAR(45)),
... Column('last_name', VARCHAR(45)))
>>> user_table
Table('user', MetaData(bind=None), Column...)
# CREATE TABLE by crete()
>>> user_table.create(bind=e)
# DROP TABLE with drop()
>>> user_table.drop(bind=e)
defining table

78. Table/Column vs table/column
Visitable (base type)
ClauseElement
Selectable
FromClause
[ table() ]
ColumnElement
ColumnClause
[ column() ]
Column Table
SchemaItem

79. Table/Column vs table/column
Visitable (base type)
ClauseElement
Selectable
FromClause
[ table() ]
ColumnElement
ColumnClause
[ column() ]
Column Table
SchemaItem
Supprorts
same operation

80. HANDS ON: Schema definition
• define "user" table with
Table()/Column()
• create table with .create()
• insert records
• drop table with .drop()
name type options
id INTEGER primary_key
username
VARCHAR
(64)
nullable=False
email
VARCHAR
(128)
nullable=False

81. ORM basics

82. Object-Relational Mapping
• Map DB records to objects
• ActiveRecord pattern
• 1 table ~> 1 class, 1 record ~> 1 object
• FK reference -> reference to other object
• column -> property/attribute of object

83. ORM in SQLAlchemy
• mapper maps a table with Python class
mapped class will have instrumented attribute
• Session manipulates DB for you to retrieve / save
mapped objects

84. mapping
Table
Column foo
Column bar
Column baz
...
entity class
"mapped" entity class
instrumented foo
instrumented bar
instrumented baz
...
method qux
method quux
method qux
method quux
...
mapper

85. classic mapping
[new] declarative mapping
mapping patterns

86. >>> from sqlalchemy.orm import mapper
>>> actor_tbl = Table(...)
>>> class Actor(object): pass
>>> mapper(Actor, actor_tbl)
>>> dir(Actor)
['__class__', ... '_sa_class_manager',
'actor_id', 'first_name', 'last_name']
>>> Actor.actor_id
<...InstrumentedAttribute object at ...>
classic mapping

87. >>> from sqlalchemy.ext.declarative import
declarative_base
>>> from sqlalchmy import DateTime, Integer, String
>>> Base = declarative_base()
>>>
>>> class Actor(Base):
... __tablename__ = 'actor'
... actor_id = Column(Integer, primary_key=True)
... first_name = Column(String)
... last_name = Column(String)
... last_update = Column(DateTime)
declarative mapping

88. >>> from sqlalchemy.ext.declarative import declarative_base
>>> from sqlalchmy import DateTime, Integer, String
>>> Base = declarative_base()
>>>
>>> class Actor(Base):
... __tablename__ = 'actor'
... actor_id = Column(Integer, primary_key=True)
... first_name = Column(String)
... last_name = Column(String)
... last_update = Column(DateTime)
>>> dir(Actor)
['__class__', '_decl_class_registry', '_sa_class_manager',
'actor_id', 'first_name', 'last_name', 'last_update',
'metadata']
declarative mapping

89. ORM in SQLAlchemy
• mapper maps a table with Python class
mapped class will have instrumented attribute
• Session manipulates DB for you to retrieve / save
mapped objects

90. session
Engine
Program
query A
select A
record Aobject B
start
tracking A...
(updates) flag A as
"dirty"
reflect new/dirty
changes
flush
start
tracking B
...
add
object B
update A
insert B
commit
begin

91. Using Session
>>> from sqlalchemy.orm import sessionmaker
>>> Session = sessionmaker(bind=e)
>>> session = Session()
>>>
>>>

92. Using Session
>>> from sqlalchemy.orm import sessionmaker
>>> Session = sessionmaker(bind=e)
>>> session = Session()
>>> query = session.query(Actor)
>>> query
<sqlalchemy.orm.query.Query object at ...>

93. Using Session
>>> from sqlalchemy.orm import sessionmaker
>>> Session = sessionmaker(bind=e)
>>> session = Session()
>>> query = session.query(Actor)
>>> query
<sqlalchemy.orm.query.Query object at ...>
>>> actor = query.first()
>>> actor
<...Actor Object at ...>

94. Query object methods:
query filtering
>>> q = session.query(Actor)
>>> str(q)
>>> 'SELECT ... nFROM actor'
>>> q1 = q.filter(Actor.first_name=='BEN')
>>> str(q1)
'SELECT ... nFROM actornWHERE actor.first_name = ?'
>>> q2 = q1.filter_by(last_name='SMITH')
>>> str(q2)
'SELECT ... nFROM actornWHERE actor.first_name = ?
AND actor.last_name= ?'

95. Query object methods:
fetching record(s)
>>> q = session.query(Actor)
>>> q.first()
<...Actor object at ...>
>>> q.all() # WARNING: SELECTs all records
<...Actor object at ...>, <...Actor object at ...>, ...
>>> q[:3]
<...Actor object at ...>, <...Actor object at ...>, ...
>>> q.count()
200

96. Update/Insert in Session
>>> session = Session()
>>> actor_a = Actor(first_name='KEN',
... last_name='WATANABE')
>>> session.add(actor_a)
>>> session.commit()
>>> actor_b = session.query(Actor).get(1)
>>> actor_b.last_name='GEORGE'
>>> session.commit()

97. HANDS ON: ORM basics
• define Category model with
declarative ORM
• select any record
• add "Nature" category
• delete "Nature" category