2023-12-05 18:03:35 +01:00
# Flow control: match case
{:.no_toc}
< nav markdown = "1" class = "toc-class" >
* TOC
{:toc}
< / nav >
## The goal
There is a new flow control in town. A switch case replacement, called match case.
Questions to [David Rotermund ](mailto:davrot@uni-bremen.de )
## The [match](https://docs.python.org/3/reference/compound_stmts.html#the-match-statement) statement
The match statement is used for pattern matching. Syntax:
```python
match_stmt ::= 'match' subject_expr ":" NEWLINE INDENT case_block+ DEDENT
subject_expr ::= star_named_expression "," star_named_expressions?
| named_expression
case_block ::= 'case' patterns [guard] ":" block
```
## Normal "switch" case
```python
for i in range(0, 4):
match (i):
case 0:
print("This is a 0")
case 0:
print("This is a 0 too.")
case 1:
print("This is a 1.")
case 2:
print("This is a 2.")
case _:
print(f"I don't know what to do with a {i}!")
```
Output:
```python
This is a 0
This is a 1.
This is a 2.
I don't know what to do with a 3!
```
\_ is the default. And after every case state has an in-build **break;** (If you don't know what this means: A. You are not a C/C++ programmer. B. You can ignore this information. :-) ).
## Unpacking parameters
2023-12-07 00:07:51 +01:00
{: .topic-optional}
This is an optional topic!
2023-12-05 18:03:35 +01:00
Different from a normal switch we can deal with several input parameters at the same time and even use wildcards:
```python
def what_to_do(input):
match (input):
case (0, 0):
print("This is (0,0) - A")
case (0, 1):
print("This is (0,1) - B")
case (x, 0):
print(f"This is a {x} - C")
case (x, y):
print(f"This is a {x} and {y} - D")
case _:
print(f"I don't know what to do with {input}- E!")
print("")
second_parameter: int = 0
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
second_parameter = 1
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
input = "Hello!"
print(f"Input: {input}")
what_to_do(input)
```
Output:
```python
Input: (0, 0)
This is (0,0) - A
Input: (1, 0)
This is a 1 - C
Input: (0, 1)
This is (0,1) - B
Input: (1, 1)
This is a 1 and 1 - D
Input: Hello!
I don't know what to do with Hello!- E!
```
2023-12-07 00:07:51 +01:00
### OR
{: .topic-optional}
This is an optional topic!
2023-12-05 18:03:35 +01:00
We can combine patter by | (or):
```python
def what_to_do(input):
match (input):
case (0, 0) | (0, 1):
print("This is (0,0) or (0,1) - A")
case (x, 0):
print(f"This is a {x} - C")
case (x, y):
print(f"This is a {x} and {y} - D")
case _:
print(f"I don't know what to do with {input}- E!")
print("")
second_parameter: int = 0
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
second_parameter = 1
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
```
Output:
```python
Input: (0, 0)
This is (0,0) or (0,1) - A
Input: (1, 0)
This is a 1 - C
Input: (0, 1)
This is (0,0) or (0,1) - A
Input: (1, 1)
This is a 1 and 1 - D
```
You can also use the **or** on parts of the pattern (output is the same as before):
```python
def what_to_do(input):
match (input):
case (0, (0 | 1)):
print("This is (0,0) or (0,1) - A")
case (x, 0):
print(f"This is a {x} - C")
case (x, y):
print(f"This is a {x} and {y} - D")
case _:
print(f"I don't know what to do with {input}- E!")
print("")
second_parameter: int = 0
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
second_parameter = 1
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
```
2023-12-07 00:07:51 +01:00
### If guards
{: .topic-optional}
This is an optional topic!
2023-12-05 18:03:35 +01:00
We can distinguish between cases by using **if guard** :
```python
def what_to_do(input):
match (input):
case (x, y) if x == y:
print(f"Both are {x} - A ")
case (x, y):
print(f"This is a {x} and {y} - B ")
print("")
second_parameter: int = 0
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
second_parameter = 1
for i in range(0, 2):
input = (i, second_parameter)
print(f"Input: {input}")
what_to_do(input)
```
Output:
```python
Input: (0, 0)
Both are 0 - A
Input: (1, 0)
This is a 1 and 0 - B
Input: (0, 1)
This is a 0 and 1 - B
Input: (1, 1)
Both are 1 - A
```
2023-12-07 00:07:51 +01:00
### Data classes
{: .topic-optional}
This is an optional topic!
2023-12-05 18:03:35 +01:00
We can combine it nicely with data classes:
```python
from dataclasses import dataclass
@dataclass
class SomeDataStructure:
x: int
y: int
def what_to_do(input):
print(input)
match (input):
case SomeDataStructure(x=0, y=0):
print("case A")
case SomeDataStructure(x=1, y=0):
print("case B")
case SomeDataStructure():
print("case C")
case _:
print("What?")
print("")
data_0_0 = SomeDataStructure(0, 0)
what_to_do(data_0_0)
data_0_1 = SomeDataStructure(0, 1)
what_to_do(data_0_1)
data_1_0 = SomeDataStructure(1, 0)
what_to_do(data_1_0)
data_1_1 = SomeDataStructure(1, 1)
what_to_do(data_1_1)
data_broken = "I am broken!"
what_to_do(data_broken)
```
Output:
```python
SomeDataStructure(x=0, y=0)
case A
SomeDataStructure(x=0, y=1)
case C
SomeDataStructure(x=1, y=0)
case B
SomeDataStructure(x=1, y=1)
case C
I am broken!
What?
```
2023-12-07 00:07:51 +01:00
### As
{: .topic-optional}
This is an optional topic!
2023-12-05 18:03:35 +01:00
We can use **as** to name a variable in a case clause (if you have a sequence i.e. tuple or list of elements then you can use **as** on the individual elements):
```python
from dataclasses import dataclass
@dataclass
class SomeDataStructure:
x: int
y: int
def what_to_do(input):
print(input)
match (input):
case SomeDataStructure(x=0, y=0):
print("case A")
case SomeDataStructure(x=1, y=0):
print("case B")
case SomeDataStructure() as someinput:
print(f"case C : {someinput.x} {someinput.y}")
case _:
print("What?")
print("")
data_0_0 = SomeDataStructure(0, 0)
what_to_do(data_0_0)
data_0_1 = SomeDataStructure(0, 1)
what_to_do(data_0_1)
data_1_0 = SomeDataStructure(1, 0)
what_to_do(data_1_0)
data_1_1 = SomeDataStructure(1, 1)
what_to_do(data_1_1)
data_broken = "I am broken!"
what_to_do(data_broken)
```
Output:
```python
SomeDataStructure(x=0, y=0)
case A
SomeDataStructure(x=0, y=1)
case C : 0 1
SomeDataStructure(x=1, y=0)
case B
SomeDataStructure(x=1, y=1)
case C : 1 1
I am broken!
What?
```
2023-12-07 00:07:51 +01:00
### List
{: .topic-optional}
This is an optional topic!
2023-12-05 18:03:35 +01:00
This is how we can deal with lists that have different length:
```python
def what_to_do(input):
print(input)
match (input):
case []:
print("case A")
case [0]:
print("case B")
case [x]:
print(f"case C {x}")
case [0, 0]:
print("case D")
case _:
print("What?")
print("")
what_to_do([])
what_to_do([0])
what_to_do([1])
what_to_do([0, 0])
what_to_do([0, 1])
what_to_do("Hello")
```
Output:
```python
[]
case A
[0]
case B
[1]
case C 1
[0, 0]
case D
[0, 1]
What?
Hello
What?
```
2023-12-07 00:07:51 +01:00
### Dictionaries
{: .topic-optional}
This is an optional topic!
2023-12-05 18:03:35 +01:00
The whole match apparatus works also with dictionaries. You can even distinguish between what kind of variable is stored behind a key (e.g. int vs str):
```python
def what_to_do(input):
print(input)
match (input):
case {"x": str() as x}:
print(f"{x} - str")
case {"x": int() as x}:
print(f"{x} - int")
case {"y": value}:
print(f"{value} - A")
case {"a": value_a, "b": value_b}:
print(f"a={value_a} b={value_b} - B")
print("")
dictionary_a = {"x": str(1)}
what_to_do(dictionary_a)
dictionary_b = {"x": int(1)}
what_to_do(dictionary_b)
dictionary_c = {"x": int(1), "y": 5}
what_to_do(dictionary_c)
dictionary_d = {"y": 5}
what_to_do(dictionary_d)
dictionary_e = {"a": 6, "b": 7}
what_to_do(dictionary_e)
dictionary_f = {"b": 8, "a": 10}
what_to_do(dictionary_f)
```
Output:
```python
{'x': '1'}
1 - str
{'x': 1}
1 - int
{'x': 1, 'y': 5}
1 - int
{'y': 5}
5 - A
{'a': 6, 'b': 7}
a=6 b=7 - B
{'b': 8, 'a': 10}
a=10 b=8 - B
```
## Reference
* [PEP 636 – Structural Pattern Matching: Tutorial ](https://peps.python.org/pep-0636/ )