functions
DESCRIPTION
Python - FunctionsTRANSCRIPT
Functions
Python
Functions
Copyright © Software Carpentry 2010
This work is licensed under the Creative Commons Attribution License
See http://software-carpentry.org/license.html for more information.
A programming language should not include
everything anyone might ever want
Python Functions
A programming language should not include
everything anyone might ever want
Instead, it should make it easy for people to createInstead, it should make it easy for people to create
what they need to solve specific problems
Python Functions
A programming language should not include
everything anyone might ever want
Instead, it should make it easy for people to createInstead, it should make it easy for people to create
what they need to solve specific problems
Define functions to create higher-level operations
Python Functions
A programming language should not include
everything anyone might ever want
Instead, it should make it easy for people to createInstead, it should make it easy for people to create
what they need to solve specific problems
Define functions to create higher-level operations
"Create a language in which the solution to your
original problem is trivial."
Python Functions
original problem is trivial."
Define functions using def
Python Functions
Define functions using def
defdefdefdef greet():
returnreturnreturnreturn 'Good evening, master'
Python Functions
Define functions using def
defdefdefdef greet():
returnreturnreturnreturn 'Good evening, master'
temp = greet()
printprintprintprint temp
Good evening, master
Python Functions
Give them parameters
Python Functions
Give them parameters
defdefdefdef greet(name):
answer = 'Hello, ' + name
returnreturnreturnreturn answer
Python Functions
Give them parameters
defdefdefdef greet(name):
answer = 'Hello, ' + name
temp
'doctor'
returnreturnreturnreturn answer
temp = 'doctor'
Python Functions
stackstackstackstack valuevaluevaluevalue
Give them parameters
defdefdefdef greet(name):
answer = 'Hello, ' + name
temp
'doctor'
name
returnreturnreturnreturn answer
temp = 'doctor'
result = greet(temp)
Python Functions
stackstackstackstack valuevaluevaluevalue
Give them parameters
defdefdefdef greet(name):
answer = 'Hello, ' + name
temp
'doctor'
'Hello, doctor'
name
answer
returnreturnreturnreturn answer
temp = 'doctor'
result = greet(temp)
Python Functions
stackstackstackstack valuevaluevaluevalue
Give them parameters
defdefdefdef greet(name):
answer = 'Hello, ' + name
temp
'doctor'
'Hello, doctor'
returnreturnreturnreturn answer
temp = 'doctor'
result = greet(temp)
Python Functions
stackstackstackstack valuevaluevaluevalue
result
Each function call creates a new stack frame
Python Functions
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
Python Functions
stackstackstackstack valuevaluevaluevalue
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
val = 10
10
val
Python Functions
val = 10
stackstackstackstack valuevaluevaluevalue
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
val = 10
10
c
valdouble
double
double
double
Python Functions
val = 10
result = double(val)
stackstackstackstack valuevaluevaluevalue
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
val = 10
10
a
c
valdouble
double
double
double
add
add
add
add
Python Functions
val = 10
result = double(val)
stackstackstackstack valuevaluevaluevalue
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
val = 10
10
11
a
b
c
valdouble
double
double
double
add
add
add
add
Python Functions
val = 10
result = double(val)
printprintprintprint resultstackstackstackstack valuevaluevaluevalue
result
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
val = 10
10
22
c
d
valdouble
double
double
double
Python Functions
val = 10
result = double(val)
printprintprintprint resultstackstackstackstack valuevaluevaluevalue
result
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
val = 10
10
22val
Python Functions
val = 10
result = double(val)
printprintprintprint resultstackstackstackstack valuevaluevaluevalue
result
Each function call creates a new stack frame
defdefdefdef add(a):
b = a + 1b = a + 1
returnreturnreturnreturn b
defdefdefdef double(c):
d = 2 * add(c)
returnreturnreturnreturn d
val = 10
10
22val
Python Functions
val = 10
result = double(val)
printprintprintprint result
22stackstackstackstack valuevaluevaluevalue
result
Only see variables in the current and global frames
Python Functions
Only see variables in the current and global frames
Current beats global
Python Functions
Only see variables in the current and global frames
Current beats global
defdefdefdef greet(name):defdefdefdef greet(name):
temp = 'Hello, ' + name
returnreturnreturnreturn temp
temp = 'doctor'
result = greet(temp)
Python Functions
Only see variables in the current and global frames
Current beats global
defdefdefdef greet(name):defdefdefdef greet(name):
temp = 'Hello, ' + name
returnreturnreturnreturn temp
temp = 'doctor'
result = greet(temp) temp
'Hello, doctor'
'doctor'
global
global
global
global
greet
greet
greet
greet
name
temp
Python Functions
stackstackstackstack valuevaluevaluevalue
global
global
global
global
Only see variables in the current and global frames
Current beats global
defdefdefdef greet(name):defdefdefdef greet(name):
temp = 'Hello, ' + name
returnreturnreturnreturn temp
temp = 'doctor'
result = greet(temp)
printprintprintprint resulttemp
'Hello, doctor'
'doctor'
global
global
global
global
Python Functions
Hello, doctor
stackstackstackstack valuevaluevaluevalue
resultglobal
global
global
global
Can pass values in and accept results directly
Python Functions
Can pass values in and accept results directly
defdefdefdef greet(name):
returnreturnreturnreturn 'Hello, ' + name
printprintprintprint greet('doctor')
Python Functions
Can pass values in and accept results directly
defdefdefdef greet(name):
returnreturnreturnreturn 'Hello, ' + name
printprintprintprint greet('doctor')
_x1_
'doctor'
'Hello, doctor'
name
_x2_
Python Functions
stackstackstackstack valuevaluevaluevalue
Can return at any time
Python Functions
Can return at any time
defdefdefdef sign(num):
ifififif num > 0:
returnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
else:else:else:else:
return return return return -1
Python Functions
Can return at any time
defdefdefdef sign(num):
ifififif num > 0:
returnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
else:else:else:else:
return return return return -1
printprintprintprint sign(3)
Python Functions
1
Can return at any time
defdefdefdef sign(num):
ifififif num > 0:
returnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
else:else:else:else:
return return return return -1
printprintprintprint sign(3)
Python Functions
1
printprintprintprint sign(-9)
-1
Can return at any time
defdefdefdef sign(num):
ifififif num > 0:Over-use makes functions
hard to understandreturnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
else:else:else:else:
return return return return -1
printprintprintprint sign(3)
hard to understand
Python Functions
1
printprintprintprint sign(-9)
-1
Can return at any time
defdefdefdef sign(num):
ifififif num > 0:Over-use makes functions
hard to understandreturnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
else:else:else:else:
return return return return -1
printprintprintprint sign(3)
hard to understand
No prescription possible, but:
Python Functions
1
printprintprintprint sign(-9)
-1
Can return at any time
defdefdefdef sign(num):
ifififif num > 0:Over-use makes functions
hard to understandreturnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
else:else:else:else:
return return return return -1
printprintprintprint sign(3)
hard to understand
No prescription possible, but:
– a few at the beginning
to handle special cases
Python Functions
1
printprintprintprint sign(-9)
-1
Can return at any time
defdefdefdef sign(num):
ifififif num > 0:Over-use makes functions
hard to understandreturnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
else:else:else:else:
return return return return -1
printprintprintprint sign(3)
hard to understand
No prescription possible, but:
– a few at the beginning
to handle special cases
– one at the end for the
Python Functions
1
printprintprintprint sign(-9)
-1
– one at the end for the
"general" result
Every function returns something
Python Functions
Every function returns something
defdefdefdef sign(num):
ifififif num > 0:
returnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
# else:else:else:else:
# return return return return -1
Python Functions
Every function returns something
defdefdefdef sign(num):
ifififif num > 0:
returnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
# else:else:else:else:
# return return return return -1
printprintprintprint sign(3)
Python Functions
1
Every function returns something
defdefdefdef sign(num):
ifififif num > 0:
returnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
# else:else:else:else:
# return return return return -1
printprintprintprint sign(3)
Python Functions
1
printprintprintprint sign(-9)
None
Every function returns something
defdefdefdef sign(num):
ifififif num > 0:If the function doesn't return
a value, Python returns Nonereturnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
# else:else:else:else:
# return return return return -1
printprintprintprint sign(3)
a value, Python returns None
Python Functions
1
printprintprintprint sign(-9)
None
Every function returns something
defdefdefdef sign(num):
ifififif num > 0:If the function doesn't return
a value, Python returns Nonereturnreturnreturnreturn 1
elifelifelifelif num == 0:
returnreturnreturnreturn 0
# else:else:else:else:
# return return return return -1
printprintprintprint sign(3)
a value, Python returns None
Yet another reason why
commenting out blocks of code
is a bad idea...
Python Functions
1
printprintprintprint sign(-9)
None
Functions and parameters don't have types
Python Functions
Functions and parameters don't have types
defdefdefdef double(x):
returnreturnreturnreturn 2 * x
Python Functions
Functions and parameters don't have types
defdefdefdef double(x):
returnreturnreturnreturn 2 * x
printprintprintprint double(2)
4
Python Functions
Functions and parameters don't have types
defdefdefdef double(x):
returnreturnreturnreturn 2 * x
printprintprintprint double(2)
4
printprintprintprint double('two')
twotwo
Python Functions
Functions and parameters don't have types
defdefdefdef double(x):
returnreturnreturnreturn 2 * xOnly use this when the
function's behavior dependsprintprintprintprint double(2)
4
printprintprintprint double('two')
twotwo
function's behavior depends
only on properties that all
possible arguments share
Python Functions
Functions and parameters don't have types
defdefdefdef double(x):
returnreturnreturnreturn 2 * xOnly use this when the
function's behavior dependsprintprintprintprint double(2)
4
printprintprintprint double('two')
twotwo
function's behavior depends
only on properties that all
possible arguments share
ifififif type(arg) == int:
Python Functions
ifififif type(arg) == int:
...
elif typeelif typeelif typeelif type(arg) == str:
...
...
Functions and parameters don't have types
defdefdefdef double(x):
returnreturnreturnreturn 2 * xOnly use this when the
function's behavior dependsprintprintprintprint double(2)
4
printprintprintprint double('two')
twotwo
function's behavior depends
only on properties that all
possible arguments share
Warning sign ifififif type(arg) == int:
Python Functions
Warning sign ifififif type(arg) == int:
...
elif typeelif typeelif typeelif type(arg) == str:
...
...
Functions and parameters don't have types
defdefdefdef double(x):
returnreturnreturnreturn 2 * xOnly use this when the
function's behavior dependsprintprintprintprint double(2)
4
printprintprintprint double('two')
twotwo
function's behavior depends
only on properties that all
possible arguments share
Warning sign ifififif type(arg) == int:
Python Functions
Warning sign
There's a better
way to do this
ifififif type(arg) == int:
...
elif typeelif typeelif typeelif type(arg) == str:
...
...
Values are copied into parameters
Python Functions
Values are copied into parameters
Which means lists are aliased
Python Functions
Values are copied into parameters
Which means lists are aliased
defdefdefdef appender(a_string, a_list):defdefdefdef appender(a_string, a_list):
a_string += 'turing'
a_list.append('turing')
Python Functions
Values are copied into parameters
Which means lists are aliased
defdefdefdef appender(a_string, a_list):defdefdefdef appender(a_string, a_list):
a_string += 'turing'
a_list.append('turing')
string_val = 'alan'
list_val = ['alan']
appender(string_val, list_val)
Python Functions
Values are copied into parameters
Which means lists are aliased
defdefdefdef appender(a_string, a_list):defdefdefdef appender(a_string, a_list):
a_string += 'turing'
a_list.append('turing')
string_val = 'alan'
list_val = ['alan']
appender(string_val, list_val)string_val
'alan'
['alan']
Python Functions
stackstackstackstack valuevaluevaluevalue
list_val
Values are copied into parameters
Which means lists are aliased
defdefdefdef appender(a_string, a_list):defdefdefdef appender(a_string, a_list):
a_string += 'turing'
a_list.append('turing')
string_val = 'alan'
list_val = ['alan']
appender(string_val, list_val)string_val
'alan'
['alan']
a_string
a_list
Python Functions
stackstackstackstack valuevaluevaluevalue
list_val
Values are copied into parameters
Which means lists are aliased
defdefdefdef appender(a_string, a_list):defdefdefdef appender(a_string, a_list):
a_string += 'turing'
a_list.append('turing')
string_val = 'alan'
list_val = ['alan']
appender(string_val, list_val)string_val
'alan'
'alanturing'
['alan']
a_string
a_list
Python Functions
stackstackstackstack valuevaluevaluevalue
list_val
Values are copied into parameters
Which means lists are aliased
defdefdefdef appender(a_string, a_list):defdefdefdef appender(a_string, a_list):
a_string += 'turing'
a_list.append('turing')
string_val = 'alan'
list_val = ['alan']
appender(string_val, list_val)string_val
'alan'
'alanturing'
['alan',
a_string
a_list
Python Functions
stackstackstackstack valuevaluevaluevalue
list_val 'turing']
Values are copied into parameters
Which means lists are aliased
defdefdefdef appender(a_string, a_list):defdefdefdef appender(a_string, a_list):
a_string += 'turing'
a_list.append('turing')
string_val = 'alan'
list_val = ['alan']
appender(string_val, list_val)
printprintprintprint string_val
string_val
'alan'
['alan',
Python Functions
printprintprintprint string_val
alan
printprintprintprint list_val
['alan', 'turing']stackstackstackstack valuevaluevaluevalue
list_val 'turing']
Can define default parameter values
Python Functions
Can define default parameter values
defdefdefdef adjust(value, amount=2.0):
returnreturnreturnreturn value * amount
Python Functions
Can define default parameter values
defdefdefdef adjust(value, amount=2.0):
returnreturnreturnreturn value * amount
printprintprintprint adjust(5)
10
Python Functions
Can define default parameter values
defdefdefdef adjust(value, amount=2.0):
returnreturnreturnreturn value * amount
printprintprintprint adjust(5)
10
printprintprintprint adjust(5, 1.001)
5.005
Python Functions
More readable than multiple functions
Python Functions
More readable than multiple functions
defdefdefdef adjust_general(value, amount):
returnreturnreturnreturn value * amount
defdefdefdef adjust_default(value):
returnreturnreturnreturn adjust_general(value, 2.0)
Python Functions
Parameters that have defaults must come after
parameters that do not
Python Functions
Parameters that have defaults must come after
parameters that do not
defdefdefdef triplet(left='venus', middle, right='mars'):
returnreturnreturnreturn '%s %s %s' % (left, middle, right)
Python Functions
Parameters that have defaults must come after
parameters that do not
defdefdefdef triplet(left='venus', middle, right='mars'):
returnreturnreturnreturn '%s %s %s' % (left, middle, right)
printprintprintprint triplet('earth')
venus earth mars
OK so far...
Python Functions
Parameters that have defaults must come after
parameters that do not
defdefdefdef triplet(left='venus', middle, right='mars'):
returnreturnreturnreturn '%s %s %s' % (left, middle, right)
printprintprintprint triplet('earth')
venus earth mars
printprintprintprint triplet('pluto', 'earth')
OK so far...
?
Python Functions
printprintprintprint triplet('pluto', 'earth') ?
Parameters that have defaults must come after
parameters that do not
defdefdefdef triplet(left='venus', middle, right='mars'):
returnreturnreturnreturn '%s %s %s' % (left, middle, right)
printprintprintprint triplet('earth')
venus earth mars
printprintprintprint triplet('pluto', 'earth')
OK so far...
?
Python Functions
printprintprintprint triplet('pluto', 'earth')
triplet('pluto', 'earth', 'mars')
?
Parameters that have defaults must come after
parameters that do not
defdefdefdef triplet(left='venus', middle, right='mars'):
returnreturnreturnreturn '%s %s %s' % (left, middle, right)
printprintprintprint triplet('earth')
venus earth mars
printprintprintprint triplet('pluto', 'earth')
OK so far...
?
Python Functions
printprintprintprint triplet('pluto', 'earth')
triplet('venus', 'pluto', 'earth')
triplet('pluto', 'earth', 'mars')
?
"When should I write a function?"
Python Functions
"When should I write a function?"
Human short term memory can hold 7± 2 items
Python Functions
"When should I write a function?"
Human short term memory can hold 7± 2 items
If someone has to keep more than a dozen thingsIf someone has to keep more than a dozen things
in their mind at once to understand a block of code,
it's too long
Python Functions
"When should I write a function?"
Human short term memory can hold 7± 2 items
If someone has to keep more than a dozen thingsIf someone has to keep more than a dozen things
in their mind at once to understand a block of code,
it's too long
Break it into comprehensible pieces with functions
Python Functions
"When should I write a function?"
Human short term memory can hold 7± 2 items
If someone has to keep more than a dozen thingsIf someone has to keep more than a dozen things
in their mind at once to understand a block of code,
it's too long
Break it into comprehensible pieces with functions
Even if each function is only called once
Python Functions
Even if each function is only called once
Example
forforforfor x inininin range(1, GRID_WIDTH-1):
forforforfor y inininin range(1, GRID_HEIGHT-1):
ifififif (density[x-1][y] > density_threshold) or \
(density[x+1][y] > density_threshold):
ifififif (flow[x][y-1] < flow_threshold) or\
(flow[x][y+1] < flow_threshold):
temp = (density[x-1][y] + density[x+1][y]) / 2
ifififif abs(temp - density[x][y]) > update_threshold:
density[x][y] = temp
Python Functions
Refactoring #1: grid interior
forforforfor x inininin grid_interior(GRID_WIDTH):
forforforfor y inininin grid_interior(GRID_HEIGHT):
ifififif (density[x-1][y] > density_threshold) or \
(density[x+1][y] > density_threshold):
ifififif (flow[x][y-1] > flow_threshold) or\
(flow[x][y+1] > flow_threshold):
temp = (density[x-1][y] + density[x+1][y]) / 2
ifififif abs(temp - density[x][y]) > update_threshold:
density[x][y] = temp
Python Functions
Refactoring #2: tests on X and Y axes
forforforfor x inininin grid_interior(GRID_WIDTH):
forforforfor y inininin grid_interior(GRID_HEIGHT):
ifififif density_exceeds(density, x, y, density_threshold):
ifififif flow_exceeds(flow, x, y, flow_threshold):
temp = (density[x-1][y] + density[x+1][y]) / 2
ifififif abs(temp - density[x][y]) > tolerance:
density[x][y] = temp
Python Functions
Refactoring #3: update rule
forforforfor x inininin grid_interior(GRID_WIDTH):
forforforfor y inininin grid_interior(GRID_HEIGHT):
ifififif density_exceeds(density, x, y, density_threshold):
ifififif flow_exceeds(flow, x, y, flow_threshold):
update_on_tolerance(density, x, y, tolerance)
Python Functions
Refactoring #3: update rule
forforforfor x inininin grid_interior(GRID_WIDTH):
forforforfor y inininin grid_interior(GRID_HEIGHT):
ifififif density_exceeds(density, x, y, density_threshold):
ifififif flow_exceeds(flow, x, y, flow_threshold):
update_on_tolerance(density, x, y, tolerance)
Good programmers will write this first
Python Functions
Refactoring #3: update rule
forforforfor x inininin grid_interior(GRID_WIDTH):
forforforfor y inininin grid_interior(GRID_HEIGHT):
Good programmers will write this first
Then write the functions it implies
ifififif density_exceeds(density, x, y, density_threshold):
ifififif flow_exceeds(flow, x, y, flow_threshold):
update_on_tolerance(density, x, y, tolerance)
Python Functions
Refactoring #3: update rule
forforforfor x inininin grid_interior(GRID_WIDTH):
forforforfor y inininin grid_interior(GRID_HEIGHT):
Good programmers will write this first
Then write the functions it implies
ifififif density_exceeds(density, x, y, density_threshold):
ifififif flow_exceeds(flow, x, y, flow_threshold):
update_on_tolerance(density, x, y, tolerance)
Python Functions
Then refactor any overlap
October 2010
created by
Greg Wilson
October 2010
Copyright © Software Carpentry 2010
This work is licensed under the Creative Commons Attribution License
See http://software-carpentry.org/license.html for more information.