An introduction to Pythonapiacoa.org/publications/teaching/python/intro-python.pdf · Python and data science I Python is one of the two de facto standard languages for data science

An introduction to Python

Fabrice Rossi

CEREMADEUniversité Paris-Dauphine

2019

Python

https://www.python.org/

I Python is a high level programminglanguage

I Python’s reference implementationis a multiplatform free software

I Python can be extended bythousands of libraries

I Python is generally considered to beeasy to learn

name = input("What's your name? ")print("Hello", name)

What's your name? John DoeHello John Doe

2

https://www.python.org/

Python and data science

I Python is one of the two defacto standard languages fordata science (with R)

I Python has a large collectionof high performance datascience oriented libraries

I Python is generallyconsidered to be easy toread

3

Python

ProsI open source implementationI full-fledged programming

languageI strong support from a large

communityI broad coverage of data

science, statistics, etc.I high performance librariesI high quality graphicsI curated distribution

ConsI limited point-and-click

supportI rather steep learning curve

compared to an integratedsoftware

I naive code has lowperformances

I “old” language (1990) with alack of modern constructs

4

Recommended installs

I Anaconda (with Python 3.x)I https://www.anaconda.com/distribution/I a python distribution: python + libraries + toolsI data science orientedI anaconda navigator for managing the distribution

I recommended tools (in Anaconda)I VS code or Spyder for Python programmingI JupyterLab for literate programming

I other IDE include PyCharmI do not use Python 2.7

5

https://www.anaconda.com/distribution/

https://www.jetbrains.com/pycharm/

Outline

Introduction

Core concepts

Control structures

Functions

Exception handling

6

Outline

Introduction

Core conceptsProgramming LanguageConsole interactionBasic data modelVariablesStringsFunctionsModules

Control structures

Functions

Exception handling

7

Programming Language

DefinitionI a formal language with a strict mathematical definitionI defines syntactically correct programsI associated to a semantics

I (formal) model of the computerI effects of a program on the model

In other words...I a programming language can be used to write programs ' textsI a programming language has a strict syntax

I lexical aspects ' word spellingI grammatical aspects ' sentence level

I when a program follows the syntax, it has a proper meaning i.e.an effect on the computer on which it runs

8

Programming Language

DefinitionI a formal language with a strict mathematical definitionI defines syntactically correct programsI associated to a semantics

I (formal) model of the computerI effects of a program on the model

In other words...I a programming language can be used to write programs ' textsI a programming language has a strict syntax

I lexical aspects ' word spellingI grammatical aspects ' sentence level

I when a program follows the syntax, it has a proper meaning i.e.an effect on the computer on which it runs

8

A computer

Turing MachineI standard mathematical modelI too low level to a daily use

Other modelsI data oriented modelsI a model of the dataI together with a model of the execution of a program

I effects of instructions/statements on the data ' sentence levelI global flow and organization on a program ' text level

I include input/output aspects

9

https://en.wikipedia.org/wiki/Turing_machine

Interactive mode

Standard program executionI a program is written in a file (or a set of files)I in some languages the file can be translated to a more efficient

languageI the file (or its translation) is executed on a computer

Console/ShellI some languages have an associated “console” or “shell” (e.g.

Python and R)I one can type interactively program sentences and get associated

resultsI simplifies learning and testing

10

Python Shell

I Python provides a shell forinteractive use

I in general integrated in aspecific window of aprogramming environment

I can be launched from thecommand line (python)

I command prompt >>>

WarningThe behavior of a program in theshell is not exactly the same as thebehavior of a program outside of theshell

11

Python Shell





Python 3.7.3 (default, Mar 27 2019, 22:11:17)[GCC 7.3.0] :: Anaconda, Inc. on linuxType "help", "copyright", "credits" or "license"

for more information.>>>


11

Python Shell






for more information.>>> 2 + 2


11

Python Shell






for more information.>>> 2 + 24>>>


11

Python Shell






for more information.>>> 2 + 24>>> 4 ** 3


11

Python Shell






for more information.>>> 2 + 24>>> 4 ** 364>>>


11

Python Shell






for more information.>>> 2 + 24>>> 4 ** 364>>>


11

Python Shell as a calculator

>>> 2.5 / 1.31.923076923076923>>> 2,5 / 1,3(2, 5.0, 3)>>> 1 + 2 / 31.6666666666666665>>> (1 + 2) / 31.0>>> 5 / 22.5>>> 5 // 22>>> 5 % 21>>> -5 // 2-3>>> 5 ** 53125>>> 2 ** 0.51.4142135623730951

>>> 12.5 - 4 / 511.7>>> _ + 213.7>>> 4.5 > 3.5True>>> (2.5 >= 3) or (2.5 < 3)True>>> -1 ** 0.5

-1.0>>> (-1) ** 0.5(6.123233995736766e-17+1j)>>> _ ** 2(-1+1.2246467991473532e-16j)>>> 0j0j>>> 1j ** 2(-1+0j)

12

Basic data model

Numerical valuesI integersI real numbers

I decimal pointI classical scientific notation

e.g. 1.5e-3I complex numbers

I automatically used in somesituations

I real + img j

Arithmetic operationsI standard operationsI integer oriented

Logical expressionsI boolean (a.k.a. truth value)I True and False valuesI automatic integer conversion

to 1 and 0, respectively (andvice versa)

I logical operators and or notI numerical comparisons

I == and !=I <= and < (and reversed

ones)

13

Basic data model

Numerical valuesI integersI real numbers

I decimal pointI classical scientific notation

e.g. 1.5e-3I complex numbers

I automatically used in somesituations

I real + img j

Arithmetic operationsI standard operationsI integer oriented

Logical expressionsI boolean (a.k.a. truth value)I True and False valuesI automatic integer conversion

to 1 and 0, respectively (andvice versa)

I logical operators and or notI numerical comparisons

I == and !=I <= and < (and reversed

ones)

13

Basic data model

SyntaxI literal values (spelling)

I e.g. numbers and truthvalues

I Python specifies how towrite them

I e.g. 1,5 is not a realnumber!

I operations (grammar)I writing rules are similar to

mathematical onesI with exceptions such as

I == for equalityI ** for exponentiationI etc.

SemanticsI interpretation off the symbols

and of the expressions suchas:I calculation orderingI == tests for equalityI ** can produce complex

numbersI _ is the last value

computedI error cases

>>> 0/0Traceback (most recent call last):File "<stdin>", line 1, in <module>

ZeroDivisionError: division by zero

14

Objects and variables

ObjectsI Python manipulates objectsI each object has a type

I specifies the possiblevalues

I specifies the possibleoperations

I examplesI 2 is an intI 2.5 is a floatI True is a boolI 1+2j is a complex

VariablesI objects can be namedI a variable is a name for an

objectI setting/binding a name:

variable = object

I when a name appears in anexpression it is replaced bythe object

I example>>> x = 2>>> 2 * x4

15

Examples

>>> x = 4>>> y = 3>>> y / x0.75>>> zTraceback (most recent call last):File "<stdin>", line 1, in <module>

NameError: name 'z' is not defined>>> y / XTraceback (most recent call last):File "<stdin>", line 1, in <module>

NameError: name 'X' is not defined>>> z = x>>> z4>>> x = 3>>> z4>>> y = z < x>>> yFalse

Key pointsI (obvious) sequential modelI no default bindingI case dependantI aliases: several names for

a given objectI variable = variable

does not bind the namestogether

I unconstrained rebinding

16

Texts

>>> "A text"'A text'>>> 'Another text''Another text'>>> '''yet... another... text''''yet\nanother\ntext'>>> ''''>>> u = 'my text'>>> u * 2'my textmy text'>>> t = ' is mine!'>>> u + t'my text is mine!'>>> 2 + tTraceback (most recent call last):File "<stdin>", line 1, in <module>

TypeError: unsupported operand type(s) for +: 'int' and 'str'

StringsI type str (string)I literal ' ' or " "

I multiline with ''' '''

I concatenationI types are not compatible in

general!

17

Slices

>>> x = 'abcdefg'>>> x[0]'a'>>> x[4]'e'>>> x[-1]'g'>>> x[0:3]'abc'>>> x[:4]'abcd'>>> x[2:]'cdefg'>>> x[-3:]'efg'>>> x[:-3]'abcd'>>> x[7]Traceback (most recent call last):File "<stdin>", line 1, in <module>

IndexError: string index out of range

IndexingI some Python objects can be

indexedI [index]

I numbering always start at 0I negative indexing enables

reverse orderingI slices a:b

I from a to b-1I missing a: 0I missing b: last index + 1

I negative slicing: same logic

18

Functions

Additional actionsI objects can be manipulated with

more than operatorsI functions provide such additional

actionsI a function

I has a nameI needs 0 or more argument(s)I possibly returns an object

I using a functionI function callI function(argument_1, argument_2)

I function()

>>> len('abcd')4>>> type(2)<class 'int'>>>> type('2')<class 'str'>>>> complex(2,-1)(2-1j)>>> int(2.4)2>>> round(17.23,1)17.2>>> str(3)'3'>>> abs(-4)4>>> x = 2>>> 'x=' + str(x)'x=2'

19

Functions

Functions are objectsI type function for general

functionsI specific type for built in

functionsI all standard properties apply

I new namesI function as an argument to

a function

>>> len<built-in function len>>>> type(len)<class 'builtin_function_or_method'>>>> foo = len>>> foo<built-in function len>>>> foo('abc')3>>> str(foo)'<built-in function len>'

20

Methods

Functions for objectsI methods are specific

functions associated to someobject types

I special calling syntaxobject.function()

I equivalent toType.function(object)

>>> 'bla'.capitalize()'Bla'>>> 'tototi'.find('t')0>>> 'tototi'.find('ti')4>>> foo = 'et' * 3>>> foo'etetet'>>> foo.upper()'ETETET'>>> foo.count('et')3

21

Modules

Extending PythonI modules provide new

functions and typesI a module must be imported

to have access to its contentI default module sys

Importing modulesI import module gives

access to the names in themodule via module.name

I import module as blaturns that into bla.name

>>> import math>>> math.pi3.141592653589793>>> math.factorial(20)2432902008176640000>>> math.log(2)0.6931471805599453>>> math.ceil(3.4)4>>> import random as rd>>> rd.random()0.9786544666626154>>> rd.random()0.7496554473100112>>> rd.randint(1,10)2>>> rd.randint(1,10)8

22

Outline

Introduction

Core concepts

Control structuresNon interactive PythonConditional executionLoops

Functions

Exception handling

23

Input and output

Console limitationI has to be used interactivelyI commands are not savedI reproducibility is not

guaranteed

ScriptsI normal simple python

programs are scriptI a script: a text file (generally

ending with .py)I a script is executed by the

python interpreter

OutputsI to output something, use the

print functionI for instance

print(2, 'toto')x = 3print(x, 2 * x, 2 ** x)

will print2 toto3 6 8

InputsI to input something, use the

input functionI returns always a string str

I convert if needed

24

Conditional execution

Execute if...I programs can include parts

that are executed only ifsome condition is fulfilled

I the condition is written as aBoolean expression

General formif expression:

statement_1statement_2...statement_n

rest of the program

if in PythonI if is a compound statementI it consists a clause

comprisingI a header

if expression:I a suite whose execution is

controlled by the headerI in general the suite (a.k.a.

the body) is made of a seriesof indented statements

Semanticsthe body is executed if and only ifthe expression of the clauseevaluates to True

25

More conditional execution

Other clauses in ifI a if statement can contain

I one or more elif clausesI one else clause

I general formif expression_1:

statement_1...statement_n

elif expression_2:...

elif expresion_3:...

else:...

rest of the program

SemanticsThe compound instruction is executed asfollows

I the expression of the if header isevaluated

I if the value is True then the body isexecuted and the execution resumes forthe rest of the program

I if the value is False the body is ignoredthe execution resumes on the secondclause

I for each elif header, the execution

follows the same pattern:I if the corresponding expression

is True the body of the clause isexecuted, followed by the rest ofthe program

I if not the execution resumes onthe next clause

I if all expressions evaluate to False thebody of the else clause is executed

26

Repeating instructions

Multiple executionsI programs can include parts

that are executed severaltimes

I repetitions can be conditionalor numbered

Conditional loopwhile expression:


rest of the program

while in PythonI compound statement (single

clause)I while expression: is

the header of the clause

SemanticsI the expression of the header is

evaluatedI if the value is True

I the body is executedI the execution resumes on clause

itself!

I if the value is False theexecution resumes for the rest ofthe program

27

Iteration

Iterable objectsI objects which can be

decomposed into severalother objects

I the content of an iterableobject is arranged in acertain order

I iterating over the objectmeans accessing in order toits elements

StringsI string "content": charactersI iterating a string: in character

order!I 'foobar' gives 'f', 'o',

'o', 'b', 'a' and 'r'

28

Iterating iterables

For loopsI specific loop for iterablesI the loop execute a code for

each value contained in theiterable

General formfor variable in expression:


rest of the program

for in PythonI compound statement (single

clause)I for variable in expression: is

the header of the clauseI the expression of the header

must evaluate to an iterableobject

29

For semantics

SemanticsI the expression is evaluated

to get an iterable objectI for each object in the iterable

I the variable is bound tothe object

I the body of the clause isexecuted

I then the execution of the restof the program resumes

I if the iterable is empty, the forloop does not execute (noerror)

ExampleThe programfor x in 'foobar':

print(x)

printsfoobar

30

Ranges

Repeating n times someoperationsI very common caseI easy to do with a while but

not immediately obviousk = 0while k < n:

somethingto repeatn timesk = k + 1

more statements

range objectsI integer range iterableI range(n): integers from 0

to n-1 (n values)I simpler solution

for k in range(n):somethingto repeatn times

more statements

I clearer for pythonprogrammer

31

Ranges

More rangesI range operates in a similar way to slicesI range(end): integers from 0 to end-1

I range(begin, end): integers from begin to end-1I range(begin, end, step): integers from begin to end-1 by

increments of stepI range(1, 4, 2) : 1 and 3I range(1, 5, 2) : 1 and 3

I works with negative incrementsI range(5, 2) : emptyI range(5, 2, -1): 5, 4 and 3

32

Outline

Introduction

Core concepts

Control structures

FunctionsDefining functionsNamespacesRecursive functionsParameters and arguments

Exception handling

33

Defining functions

Benefits of user definedfunctionsI provide program organizationI reduce code repetitionI enable using generic

functionalities

Exampledef onemore(x):

return x + 1

General formdef function_name(p_1,...,p_n):


returnI the return statement

defines the value of thefunction

I it terminates the functionexecution

VocabularyI the code above is a function

definitionI p_1,...,p_n are the

formal parameters of thefunction (possible none)

I the statements form the bodyof the function

34

Calling a function

Definition versus callI the function definition only

makes it available in the restof the program

I a (standard) function call isneeded to use itfunction_name(a_1,...,a_n)

I the expressiona_1,...,a_n are thearguments of the call

Semanticsa function call is evaluated as follows

1. arguments are evaluated

2. a new namespace is created

3. formal parameters becomevariables in the new namespaceand are bound to thecorresponding arguments

4. the body of the function isexecuted

5. the namespace is discarded

6. the value of the function call isthe result of the execution of thebody

35

Example

Program1 def onemore(x):2 return x + 13

4 a = 25 b = onemore(a + 2)

Namespaces

global I onemoreI a → 2I b → 5

local I x → 4

Execution

I lines 1 and 2:I function definitionI onemore is added to the global namespaceI no other statement are executed

I line 4: a added to the global namespace withvalue 2

I line 5:

I a + 2 is evaluated to 4I a local namespace is createdI formal parameters are bound to argumentsI line 2 is executed

I x + 1 is evaluated to 5I the return value of onemore is bound to 5

I the local namespace is discardedI onemore(a + 2) is evaluated to 5I b is bound 5 in the global namespace

36

Example


4 a = 25 b = onemore(a + 2)

Namespacesglobal I onemore

I a → 2I b → 5

local I x → 4

ExecutionI lines 1 and 2:

I function definitionI onemore is added to the global namespaceI no other statement are executed


I line 5:




36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4




I line 5:




36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4




I line 5:




36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4




I line 5:I a + 2 is evaluated to 4

I a local namespace is createdI formal parameters are bound to argumentsI line 2 is executed



36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4




I line 5:I a + 2 is evaluated to 4I a local namespace is created

I formal parameters are bound to argumentsI line 2 is executed



36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4




I line 5:I a + 2 is evaluated to 4I a local namespace is createdI formal parameters are bound to arguments

I line 2 is executedI x + 1 is evaluated to 5I the return value of onemore is bound to 5


36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4




I line 5:I a + 2 is evaluated to 4I a local namespace is createdI formal parameters are bound to argumentsI line 2 is executed



36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4






I the local namespace is discarded

I onemore(a + 2) is evaluated to 5I b is bound 5 in the global namespace

36

Example


4 a = 25 b = onemore(a + 2)


I a → 2

I b → 5

local I x → 4






I the local namespace is discardedI onemore(a + 2) is evaluated to 5

I b is bound 5 in the global namespace

36

Example


4 a = 25 b = onemore(a + 2)


I a → 2I b → 5

local I x → 4







36

Return

SemanticsI return both

I binds the value of the functionI interrupts its execution

I a function can contain multiples return statements (only one willbe executed)

I when a function contains no return statementI its value is NoneI its execution continues until the end of its body

37

Examples

Multiple return1 def my_fun(x, y):2 if x > y:3 return x4 else:5 return y

I the function value is obviouslythe largest of its two arguments

I if the first argument is thelargest one, the first returnstatement is executed and thusonly lines 2 and 3 are executed

I in the other case, the secondreturn statement is executed

No return1 def foo(x):2 x = x + 13 print(x)

I lines 2 and 3 are alwaysexecuted

I the value of the function is NoneI do not confuse printing and

returning a value! The program1 def foo(x):2 x = x + 13 print(x)4

5 y = foo(2)6 print(y)

prints3None

38

Namespaces

DefinitionA namespace binds names to objects

ExamplesI the built-in namespace (with type, len, etc.)I the global namespace of a programI the local namespace of a function (during its execution)

Important aspectsI namespaces are runtime dynamical entitiesI two different namespaces can contain the same name bound to

different objects

39

Scopes

DefinitionA scope is a textual part of a program in which a namespace is directlyaccessible

ExamplesI a Python program is a scope (associated to the global namespace

of the program) which is enclosed in the scope of the built-innamespace

I a function definition defines a scope which is enclosed in theglobal scope

40

Access rules

Directly accessibleI names in the namespace of the local scope are directly accessible

(those are local names)I names in namespaces associated to enclosing function scopes

are directly accessible (when a function is defined inside anotherfunction)

I global names are accessible (names in the global enclosingnamespace)

I built-in names are accessibleI names are searched for in order from the local scope to the

built-in one: the first match is used!

41

Example

Non local accessThis program

1 x = 1 # global scope2

3 def f(y):4 # local scope of f5 return max(x, y)6

7 print(f(2))8 x = 39 print(f(2))

prints23

Do not do that!

Scopes

1. built-in2. global (the program)3. local to f

AccessesI max is accessible as a name of the built-in

namespaceI y is accessible in f as a name of the

namespace created when f is executedand attached to the scope of f

I x is accessible in f as a name of the globalnamescape attached to the global scopewhich encloses the scope of f

42

Example

Non local accessThis program

1 x = 1 # global scope2

3 def f(y):4 # local scope of f5 return max(x, y)6

7 print(f(2))8 x = 39 print(f(2))

prints23

Do not do that!

Scopes

1. built-in2. global (the program)3. local to f

AccessesI max is accessible as a name of the built-in

namespaceI y is accessible in f as a name of the

namespace created when f is executedand attached to the scope of f

I x is accessible in f as a name of the globalnamescape attached to the global scopewhich encloses the scope of f

42

Examples

Cannot access enclosedscopesIn this program

1 def f(z):2 return z + 13

4 print(f(2))5 print(z)

line 5 prints an error of the formNameError: name 'z' is not defined

z is not accessible in the globalscope.

PriorityThis program

1 def g(x):2 return x + 13

4 x = 25 print(g(3))6 print(x)

prints42

I x is both a local name(parameter) and a global one

I the name is searched first inthe local namespace andthen in enclosing ones

43

Recursive functions

Calling oneselfI a function body may

contain calls to itselfI leverage dynamic

namespaces: each callhas its own namespace

Example1 def facto(n):2 if n <= 1:3 return 14 else:5 return n * facto(n-1)

Analyzing a call

facto(4)n → 4facto(3)

n → 3facto(2)

n → 2facto(1)

n → 1return 1

n * facto(1)→ 2return 2



44

Recursive functions

Calling oneselfI a function body may

contain calls to itselfI leverage dynamic

namespaces: each callhas its own namespace

Example1 def facto(n):2 if n <= 1:3 return 14 else:5 return n * facto(n-1)

Analyzing a call

facto(4)n → 4facto(3)

n → 3facto(2)

n → 2facto(1)

n → 1return 1




44

Matching parameters and arguments

Positional matchingI standard caseI definition

def function_name(p_1,...,p_n)

I call function_name(a_1,...,a_n)I constraints and semantics

I exactly as many arguments as formalparameters

I p_k is bound to a_k

I the position of the argument decides itsformal parameters

ExampleThe programdef f(x, y):

return x - y

x = 2y = 3print(f(y, x))

prints1

45

Keyword Arguments

Name matchingI definition

def function_name(p_1,...,p_n)

I callfunction_name(p_1 = a_1,...,p_n = a_n)

I constraints and semanticsI exactly as many arguments as formal

parametersI p_k is bound to the argument associated

to its name in the callI only the names are used, not the

positionsfunction_name(p_n = a_n,...,p_1 = a_1)

ExampleThe programdef f(x, y):

return x - y

print(f(y = 3, x = 2))

prints-1

46

Mixing both types

RulesI a function call may mix

positional arguments andkeyword arguments

I positional arguments mustappear first

I when a keyword argument isused, all subsequentarguments must use thekeyword mode

ExamplesI with

def f(a, b, c):...

I incorrect callsI f(2,b=3,4)

I f(b=3,c=4,1)

I correct callsI f(2,b=3,c=4) (a is bound to

2)I f(2,c=5, b=2) (a is bound

to 2)

47

Default arguments

Main use of keywordargumentsI to enable function calls with

missing argumentsI via default values for missing

arguments

Function definitionwith default valuesdef f_n(p_1=d_1,...,p_n=d_n):


Rules and semanticsI p_k=d_k specifies both a

formal parameter p_k and itsdefault value d_k

I defaults values are optional(may be given for a subset ofthe parameters only)

I when a parameter has nomatching argument in a call,it is bound to its default value

48

Example

The program1 def foo(a, b = 2, c = 3):2 return (a + c) / b3

4 print(foo(2))5 print(foo(4, c = 2))6 print(foo(3, 5))7 print(foo(c = 4, a = 8))

prints2.53.01.26.0

Interpretationdefault values are underlined

line a b c4 2 2 35 4 2 26 3 5 37 8 2 4

49

Outline

Introduction

Core concepts

Control structures

Functions

Exception handling

50

Problems in programs

Errors and ExceptionsI Syntax errors: the program is not an acceptable python program

and cannot be executedI Exceptions: errors detected during execution

Syntax errorRunningy = 5x = 3 +/ y

printsTraceback (most recent call last):File "error.py", line 2x = 3 +/ y

^SyntaxError: invalid syntax

ExceptionRunningy = 5x = 3 +/ y

printsTraceback (most recent call last):

File "zero.py", line 3, in <module>z = x/y

ZeroDivisionError: division by zero

51

Exceptions

Handling exceptionsI normal behavior: an

exception stops theprogram

I desirable behavior: fix theproblem and continue

I mechanismI try somethingI if it does not work and

induces an exception dosomething else

Exampletry:

answer = input('Enter an integer = ')x = int(answer)

except ValueError:print(answer,' is not an integer')x = 0

print(x)

Normal outputEnter an integer = 55

Exceptional outputEnter an integer = foofoo is not an integer0

52

Handling Exceptions

try statementI try is a compound statement which starts with a try clauseI followed by

I a single finally clauseI or at least one except clause with possibly an else clause and a

finally clause

Short versiontry:

body_tfinally:body_f

Long versiontry:body_t

except type_1:body_e_1

except type_2:body_e_2...

else:body_el

finally:body_f

53

Handling Exceptions

try:body_t

except type_1:body_e_1

except type_2:body_e_2...

else:body_el

finally:body_f

rest of the program

SemanticsI Python tries to execute body_t

I if this does not produce any exception, theexecution continues through the else andthen through the rest of the program

I if an exception of type T is raisedI Python search for a matching type in the

except headers in order (an empty typein a except matches any exceptiontype)

I if a matching type is found, thecorresponding body is executed andthen the rest of the program is executed

I the finally clause is always executed, evenif an exception occurs in the except or elseclause

54

Example

try:v = input('x = ')x = int(v)print(1/x)

except ValueError:print(v,'is not an integer')

except ZeroDivisionError:print('no inverse for',x)

else:print('ok')

finally:print('this is the end')

print('rest of the program')

InteractionsI if the user inputs 2.5, she gets

2.5 is not an integerthis is the endrest of the program

I if the user inputs 4, she gets0.25okthis is the endrest of the program

I if the user inputs 0, she getsno inverse for 0this is the endrest of the program

55

Missing an exception

An exception is unhandledI when it occurs in the try

clause and is not matchedI when it occurs in a except

clauseI when it occurs in a else

clauseI when it occurs in a finally

clause

and it is passed to the enclosingenvironment

ExampleThe following programtry:

try:x = int('2.5')

except ValueError:print('got it')print(1/0)

except ZeroDivisionError:print('missed')

finally:print('exiting')

except ZeroDivisionError:print('caught')

printsgot itexitingcaught

56

Next Steps

1. Data structures in Python2. Data manipulation in Python

57

Licence

This work is licensed under a Creative CommonsAttribution-ShareAlike 4.0 International License.

http://creativecommons.org/licenses/by-sa/4.0/

58

http://creativecommons.org/licenses/by-sa/4.0/

Version

Last git commit: 2019-12-04By: Fabrice Rossi ([email protected])Git hash: 53ac9a0c6fbbf86d54b074dfebaadba23f75f6c2

59

Changelog

I November 2019: added exception handlingI October 2019: initial version

60

Documents

An introduction to Pythonapiacoa.org/publications/teaching/python/intro-python.pdf · Python and data science I Python is one of the two de facto standard languages for data science