How can I generate random integers between 0 and 9 (inclusive) in Python?
For example, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9
Try random.randrange:
from random import randrange
print(randrange(10))
Try random.randint:
import random
print(random.randint(0, 9))
Docs state:
random.randint(a, b)
Return a random integer N such that a <= N <= b.
Try this:
from random import randrange, uniform
# randrange gives you an integral value
irand = randrange(0, 10)
# uniform gives you a floating-point value
frand = uniform(0, 10)
from random import randint
x = [randint(0, 9) for p in range(0, 10)]
This generates 10 pseudorandom integers in range 0 to 9 inclusive.
The secrets module is new in Python 3.6. This is better than the random module for cryptography or security uses.
To randomly print an integer in the inclusive range 0-9:
from secrets import randbelow
print(randbelow(10))
For details, see PEP 506.
Note that it really depends on the use case. With the random module you can set a random seed, useful for pseudorandom but reproducible results, and this is not possible with the secrets module.
random module is also faster (tested on Python 3.9):
>>> timeit.timeit("random.randrange(10)", setup="import random")
0.4920286529999771
>>> timeit.timeit("secrets.randbelow(10)", setup="import secrets")
2.0670733770000425
I would try one of the following:
1.> numpy.random.randint
import numpy as np
X1 = np.random.randint(low=0, high=10, size=(15,))
print (X1)
>>> array([3, 0, 9, 0, 5, 7, 6, 9, 6, 7, 9, 6, 6, 9, 8])
2.> numpy.random.uniform
import numpy as np
X2 = np.random.uniform(low=0, high=10, size=(15,)).astype(int)
print (X2)
>>> array([8, 3, 6, 9, 1, 0, 3, 6, 3, 3, 1, 2, 4, 0, 4])
3.> numpy.random.choice
import numpy as np
X3 = np.random.choice(a=10, size=15 )
print (X3)
>>> array([1, 4, 0, 2, 5, 2, 7, 5, 0, 0, 8, 4, 4, 0, 9])
4.> random.randrange
from random import randrange
X4 = [randrange(10) for i in range(15)]
print (X4)
>>> [2, 1, 4, 1, 2, 8, 8, 6, 4, 1, 0, 5, 8, 3, 5]
5.> random.randint
from random import randint
X5 = [randint(0, 9) for i in range(0, 15)]
print (X5)
>>> [6, 2, 6, 9, 5, 3, 2, 3, 3, 4, 4, 7, 4, 9, 6]
Speed:
► np.random.uniform and np.random.randint are much faster (~10 times faster) than np.random.choice, random.randrange, random.randint .
%timeit np.random.randint(low=0, high=10, size=(15,))
>> 1.64 µs ± 7.83 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)
%timeit np.random.uniform(low=0, high=10, size=(15,)).astype(int)
>> 2.15 µs ± 38.6 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
%timeit np.random.choice(a=10, size=15 )
>> 21 µs ± 629 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)
%timeit [randrange(10) for i in range(15)]
>> 12.9 µs ± 60.4 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
%timeit [randint(0, 9) for i in range(0, 15)]
>> 20 µs ± 386 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
Notes:
1.> np.random.randint generates random integers over the half-open interval [low, high).
2.> np.random.uniform generates uniformly distributed numbers over the half-open interval [low, high).
3.> np.random.choice generates a random sample over the half-open interval [low, high) as if the argument a was np.arange(n).
4.> random.randrange(stop) generates a random number from range(start, stop, step).
5.> random.randint(a, b) returns a random integer N such that a <= N <= b.
6.> astype(int) casts the numpy array to int data type.
7.> I have chosen size = (15,). This will give you a numpy array of length = 15.
Choose the size of the array (in this example, I have chosen the size to be 20). And then, use the following:
import numpy as np
np.random.randint(10, size=(1, 20))
You can expect to see an output of the following form (different random integers will be returned each time you run it; hence you can expect the integers in the output array to differ from the example given below).
array([[1, 6, 1, 2, 8, 6, 3, 3, 2, 5, 6, 5, 0, 9, 5, 6, 4, 5, 9, 3]])
While many posts demonstrate how to get one random integer, the original question asks how to generate random integers (plural):
How can I generate random integers between 0 and 9 (inclusive) in Python?
For clarity, here we demonstrate how to get multiple random integers.
Given
>>> import random
lo = 0
hi = 10
size = 5
Code
Multiple, Random Integers
# A
>>> [lo + int(random.random() * (hi - lo)) for _ in range(size)]
[5, 6, 1, 3, 0]
# B
>>> [random.randint(lo, hi) for _ in range(size)]
[9, 7, 0, 7, 3]
# C
>>> [random.randrange(lo, hi) for _ in range(size)]
[8, 3, 6, 8, 7]
# D
>>> lst = list(range(lo, hi))
>>> random.shuffle(lst)
>>> [lst[i] for i in range(size)]
[6, 8, 2, 5, 1]
# E
>>> [random.choice(range(lo, hi)) for _ in range(size)]
[2, 1, 6, 9, 5]
Sample of Random Integers
# F
>>> random.choices(range(lo, hi), k=size)
[3, 2, 0, 8, 2]
# G
>>> random.sample(range(lo, hi), k=size)
[4, 5, 1, 2, 3]
Details
Some posts demonstrate how to natively generate multiple random integers.1 Here are some options that address the implied question:
A: random.random returns a random float in the range [0.0, 1.0)
B: random.randint returns a random integer N such that a <= N <= b
C: random.randrange alias to randint(a, b+1)
D: random.shuffle shuffles a sequence in place
E: random.choice returns a random element from the non-empty sequence
F: random.choices returns k selections from a population (with replacement, Python 3.6+)
G: random.sample returns k unique selections from a population (without replacement):2
See also R. Hettinger's talk on Chunking and Aliasing using examples from the random module.
Here is a comparison of some random functions in the Standard Library and Numpy:
| | random | numpy.random |
|-|-----------------------|----------------------------------|
|A| random() | random() |
|B| randint(low, high) | randint(low, high) |
|C| randrange(low, high) | randint(low, high) |
|D| shuffle(seq) | shuffle(seq) |
|E| choice(seq) | choice(seq) |
|F| choices(seq, k) | choice(seq, size) |
|G| sample(seq, k) | choice(seq, size, replace=False) |
You can also quickly convert one of many distributions in Numpy to a sample of random integers.3
Examples
>>> np.random.normal(loc=5, scale=10, size=size).astype(int)
array([17, 10, 3, 1, 16])
>>> np.random.poisson(lam=1, size=size).astype(int)
array([1, 3, 0, 2, 0])
>>> np.random.lognormal(mean=0.0, sigma=1.0, size=size).astype(int)
array([1, 3, 1, 5, 1])
1Namely #John Lawrence Aspden, #S T Mohammed, #SiddTheKid, #user14372, #zangw, et al.
2#prashanth mentions this module showing one integer.
3Demonstrated by #Siddharth Satpathy
You need the random python module which is part of your standard library.
Use the code...
from random import randint
num1= randint(0,9)
This will set the variable num1 to a random number between 0 and 9 inclusive.
Try this through random.shuffle
>>> import random
>>> nums = range(10)
>>> random.shuffle(nums)
>>> nums
[6, 3, 5, 4, 0, 1, 2, 9, 8, 7]
In case of continuous numbers randint or randrange are probably the best choices but if you have several distinct values in a sequence (i.e. a list) you could also use choice:
>>> import random
>>> values = list(range(10))
>>> random.choice(values)
5
choice also works for one item from a not-continuous sample:
>>> values = [1, 2, 3, 5, 7, 10]
>>> random.choice(values)
7
If you need it "cryptographically strong" there's also a secrets.choice in python 3.6 and newer:
>>> import secrets
>>> values = list(range(10))
>>> secrets.choice(values)
2
if you want to use numpy then use the following:
import numpy as np
print(np.random.randint(0,10))
>>> import random
>>> random.randrange(10)
3
>>> random.randrange(10)
1
To get a list of ten samples:
>>> [random.randrange(10) for x in range(10)]
[9, 0, 4, 0, 5, 7, 4, 3, 6, 8]
You can try importing the random module from Python and then making it choose a choice between the nine numbers. It's really basic.
import random
numbers = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
You can try putting the value the computer chose in a variable if you're going to use it later, but if not, the print function should work as such:
choice = random.choice(numbers)
print(choice)
Generating random integers between 0 and 9.
import numpy
X = numpy.random.randint(0, 10, size=10)
print(X)
Output:
[4 8 0 4 9 6 9 9 0 7]
Best way is to use import Random function
import random
print(random.sample(range(10), 10))
or without any library import:
n={}
for i in range(10):
n[i]=i
for p in range(10):
print(n.popitem()[1])
here the popitems removes and returns an arbitrary value from the dictionary n.
random.sample is another that can be used
import random
n = 1 # specify the no. of numbers
num = random.sample(range(10), n)
num[0] # is the required number
This is more of a mathematical approach but it works 100% of the time:
Let's say you want to use random.random() function to generate a number between a and b. To achieve this, just do the following:
num = (b-a)*random.random() + a;
Of course, you can generate more numbers.
From the documentation page for the random module:
Warning: The pseudo-random generators of this module should not be
used for security purposes. Use os.urandom() or SystemRandom if you
require a cryptographically secure pseudo-random number generator.
random.SystemRandom, which was introduced in Python 2.4, is considered cryptographically secure. It is still available in Python 3.7.1 which is current at time of writing.
>>> import string
>>> string.digits
'0123456789'
>>> import random
>>> random.SystemRandom().choice(string.digits)
'8'
>>> random.SystemRandom().choice(string.digits)
'1'
>>> random.SystemRandom().choice(string.digits)
'8'
>>> random.SystemRandom().choice(string.digits)
'5'
Instead of string.digits, range could be used per some of the other answers along perhaps with a comprehension. Mix and match according to your needs.
I thought I'd add to these answers with quantumrand, which uses ANU's quantum number generator. Unfortunately this requires an internet connection, but if you're concerned with "how random" the numbers are then this could be useful.
https://pypi.org/project/quantumrand/
Example
import quantumrand
number = quantumrand.randint(0, 9)
print(number)
Output: 4
The docs have a lot of different examples including dice rolls and a list picker.
I had better luck with this for Python 3.6
str_Key = ""
str_RandomKey = ""
for int_I in range(128):
str_Key = random.choice('0123456789')
str_RandomKey = str_RandomKey + str_Key
Just add characters like 'ABCD' and 'abcd' or '^!~=-><' to alter the character pool to pull from, change the range to alter the number of characters generated.
OpenTURNS allows to not only simulate the random integers but also to define the associated distribution with the UserDefined defined class.
The following simulates 12 outcomes of the distribution.
import openturns as ot
points = [[i] for i in range(10)]
distribution = ot.UserDefined(points) # By default, with equal weights.
for i in range(12):
x = distribution.getRealization()
print(i,x)
This prints:
0 [8]
1 [7]
2 [4]
3 [7]
4 [3]
5 [3]
6 [2]
7 [9]
8 [0]
9 [5]
10 [9]
11 [6]
The brackets are there becausex is a Point in 1-dimension.
It would be easier to generate the 12 outcomes in a single call to getSample:
sample = distribution.getSample(12)
would produce:
>>> print(sample)
[ v0 ]
0 : [ 3 ]
1 : [ 9 ]
2 : [ 6 ]
3 : [ 3 ]
4 : [ 2 ]
5 : [ 6 ]
6 : [ 9 ]
7 : [ 5 ]
8 : [ 9 ]
9 : [ 5 ]
10 : [ 3 ]
11 : [ 2 ]
More details on this topic are here: http://openturns.github.io/openturns/master/user_manual/_generated/openturns.UserDefined.html
Related
I want to iterate over a large list wherein I need to do some computations using n elements before the Nth index of the large list. I've solved it using the following code snippet.
mylist = [1,2,3,4,5,6,7,8,9,10,11,12,13,14]
for i in range(len(mylist)):
j=i+3
data_till_i = mylist[:j]
current_window = data_till_i[-3:]
print(current_window)
I get the following from the above code snippet:
[1, 2, 3]
[2, 3, 4]
[3, 4, 5]
[4, 5, 6]
[5, 6, 7]
[6, 7, 8]
[7, 8, 9]
[8, 9, 10]
[9, 10, 11]
[10, 11, 12]
[11, 12, 13]
[12, 13, 14]
[12, 13, 14]
[12, 13, 14]
Is there any one liner or more efficient way to do the exact same thing that will take less computation time? As my list size is very large (list has length > 100K), I'm worried about time complexity.
Thank you.
UPDATE:
My actual list is in following format:
[('string_attribute',1659675302861,3544.0), ('string_attribute', 1659675304443, 3544.0).........]
Here, the string_attribute is some attribute that is same for all the time and can be excluded from the computation.
SOLUTION:
from numpy.lib.stride_tricks import sliding_window_view
dummyList = [(1,'a'),(2,'b'),(3,'c'),(4,'d'),(5,'e'),(6,'f'),(7,'g'),(8,'h'),(9,'i'),(10,'j')]
rolling_window=sliding_window_view(dummyList, window_shape = 3, axis=0)
print(rolling_window)
The output is:
[[['1' '2' '3']
['a' 'b' 'c']]
[['2' '3' '4']
['b' 'c' 'd']]
[['3' '4' '5']
['c' 'd' 'e']]
[['4' '5' '6']
['d' 'e' 'f']]
[['5' '6' '7']
['e' 'f' 'g']]
[['6' '7' '8']
['f' 'g' 'h']]
[['7' '8' '9']
['g' 'h' 'i']]
[['8' '9' '10']
['h' 'i' 'j']]]
You can try sliding_window_view
import numpy as np
n = 3
mylist = [1,2,3,4,5,6,7,8,9,10,11,12,13,14]
window = np.lib.stride_tricks.sliding_window_view(mylist, n)
out = np.append(window, [window[-1] for _ in range(n-1)], axis=0)
print(out)
[[ 1 2 3]
[ 2 3 4]
[ 3 4 5]
[ 4 5 6]
[ 5 6 7]
[ 6 7 8]
[ 7 8 9]
[ 8 9 10]
[ 9 10 11]
[10 11 12]
[11 12 13]
[12 13 14]
[12 13 14]
[12 13 14]]
For one liner, if your Python version is greater than 3.8.0, you can try the walrus operator
out = np.append((window := np.lib.stride_tricks.sliding_window_view(mylist, n)),
[window[-1] for _ in range(n-1)], axis=0)
What you're after is called a rolling window operation. If you want to work on list type specifically, there is a shorter formulation using islice as proposed here:
window_size = 3
for i in range(len(mylist) - window_size + 1):
print(mylist[i: i + window_size])
If your data is numerical, as in the example, I'd rather propose to use numpy as this will give you much better performance! Using the proposal from here, your example becomes:
from numpy.lib.stride_tricks import sliding_window_view
sliding_window_view(np.array(mylist), window_shape = 3)
To give you a feeling for the timing, we can turn the options above into functions, create a much longer list, and compare the timing using timeit e.g. in Jupyter:
def rolling_window_using_iterator(list_, window_size):
result = []
for i in range(len(list_) - window_size + 1):
result.append(list_[i: i + window_size])
return result
def rolling_window_using_numpy(list_, window_size):
return sliding_window_view(np.array(list_), window_shape = 3)
long_list = list(range(10000000))
%timeit rolling_window_using_iterator(long_list, 3)
%timeit rolling_window_using_numpy(long_list, 3)
prints (on my machine):
1.8 s ± 22 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
422 ms ± 967 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)
I tried this way, it iterate the list in less than a second
myList = [1,2,3,4,5,6,7,8,9,10,11,12,13,14]
for index, val in enumerate(myList):
if index >= 3 :print("{} : {}".format(index, myList[index-3:index]))
The "list[index-3:index]" allow to slice the list from the nth-3 element to the nth element.
Hope it helps
List comprehension for example? (use numpy arrays for faster iteration)
import numpy as np
mylist = np.array([1,2,3,4,5,6,7,8,9,10,11,12,13,14])
chunk_size = 3
splited_list = np.array([mylist[x:x+chunk_size] for x in range(0,len(mylist)-chunk_size)])
You can cast the result to numpy array or cast every item on the list to a simple python list.
Does range function allows concatenation ? Like i want to make a range(30) & concatenate it with range(2000, 5002). So my concatenated range will be 0, 1, 2, ... 29, 2000, 2001, ... 5001
Code like this does not work on my latest python (ver: 3.3.0)
range(30) + range(2000, 5002)
You can use itertools.chain for this:
from itertools import chain
concatenated = chain(range(30), range(2000, 5002))
for i in concatenated:
...
It works for arbitrary iterables. Note that there's a difference in behavior of range() between Python 2 and 3 that you should know about: in Python 2 range returns a list, and in Python3 an iterator, which is memory-efficient, but not always desirable.
Lists can be concatenated with +, iterators cannot.
I like the most simple solutions that are possible (including efficiency). It is not always clear whether the solution is such. Anyway, the range() in Python 3 is a generator. You can wrap it to any construct that does iteration. The list() is capable of construction of a list value from any iterable. The + operator for lists does concatenation. I am using smaller values in the example:
>>> list(range(5))
[0, 1, 2, 3, 4]
>>> list(range(10, 20))
[10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
>>> list(range(5)) + list(range(10,20))
[0, 1, 2, 3, 4, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
This is what range(5) + range(10, 20) exactly did in Python 2.5 -- because range() returned a list.
In Python 3, it is only useful if you really want to construct the list. Otherwise, I recommend the Lev Levitsky's solution with itertools.chain. The documentation also shows the very straightforward implementation:
def chain(*iterables):
# chain('ABC', 'DEF') --> A B C D E F
for it in iterables:
for element in it:
yield element
The solution by Inbar Rose is fine and functionally equivalent. Anyway, my +1 goes to Lev Levitsky and to his argument about using the standard libraries. From The Zen of Python...
In the face of ambiguity, refuse the temptation to guess.
#!python3
import timeit
number = 10000
t = timeit.timeit('''\
for i in itertools.chain(range(30), range(2000, 5002)):
pass
''',
'import itertools', number=number)
print('itertools:', t/number * 1000000, 'microsec/one execution')
t = timeit.timeit('''\
for x in (i for j in (range(30), range(2000, 5002)) for i in j):
pass
''', number=number)
print('generator expression:', t/number * 1000000, 'microsec/one execution')
In my opinion, the itertools.chain is more readable. But what really is important...
itertools: 264.4522138986938 microsec/one execution
generator expression: 785.3081048010291 microsec/one execution
... it is about 3 times faster.
python >= 3.5
You can use iterable unpacking in lists (see PEP 448: Additional Unpacking Generalizations).
If you need a list,
[*range(2, 5), *range(3, 7)]
# [2, 3, 4, 3, 4, 5, 6]
This preserves order and does not remove duplicates. Or, you might want a tuple,
(*range(2, 5), *range(3, 7))
# (2, 3, 4, 3, 4, 5, 6)
... or a set,
# note that this drops duplicates
{*range(2, 5), *range(3, 7)}
# {2, 3, 4, 5, 6}
It also happens to be faster than calling itertools.chain.
from itertools import chain
%timeit list(chain(range(10000), range(5000, 20000)))
%timeit [*range(10000), *range(5000, 20000)]
738 µs ± 10.4 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
665 µs ± 13.8 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
The benefit of chain, however, is that you can pass an arbitrary list of ranges.
ranges = [range(2, 5), range(3, 7), ...]
flat = list(chain.from_iterable(ranges))
OTOH, unpacking generalisations haven't been "generalised" to arbitrary sequences, so you will still need to unpack the individual ranges yourself.
Can be done using list-comprehension.
>>> [i for j in (range(10), range(15, 20)) for i in j]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 15, 16, 17, 18, 19]
Works for your request, but it is a long answer so I will not post it here.
note: can be made into a generator for increased performance:
for x in (i for j in (range(30), range(2000, 5002)) for i in j):
# code
or even into a generator variable.
gen = (i for j in (range(30), range(2000, 5002)) for i in j)
for x in gen:
# code
With the help of the extend method, we can concatenate two lists.
>>> a = list(range(1,10))
>>> a.extend(range(100,105))
>>> a
[1, 2, 3, 4, 5, 6, 7, 8, 9, 100, 101, 102, 103, 104]
range() in Python 2.x returns a list:
>>> range(10)
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
xrange() in Python 2.x returns an iterator:
>>> xrange(10)
xrange(10)
And in Python 3 range() also returns an iterator:
>>> r = range(10)
>>> iterator = r.__iter__()
>>> iterator.__next__()
0
>>> iterator.__next__()
1
>>> iterator.__next__()
2
So it is clear that you can not concatenate iterators other by using chain() as the other guy pointed out.
You can use list function around range function to make a list
LIKE THIS
list(range(3,7))+list(range(2,9))
I came to this question because I was trying to concatenate an unknown number of ranges, that might overlap, and didn't want repeated values in the final iterator. My solution was to use set and the union operator like so:
range1 = range(1,4)
range2 = range(2,6)
concatenated = set.union(set(range1), set(range2)
for i in concatenated:
print(i)
time_interval = [4, 6, 12]
I want to sum up the numbers like [4, 4+6, 4+6+12] in order to get the list t = [4, 10, 22].
I tried the following:
t1 = time_interval[0]
t2 = time_interval[1] + t1
t3 = time_interval[2] + t2
print(t1, t2, t3) # -> 4 10 22
If you're doing much numerical work with arrays like this, I'd suggest numpy, which comes with a cumulative sum function cumsum:
import numpy as np
a = [4,6,12]
np.cumsum(a)
#array([4, 10, 22])
Numpy is often faster than pure python for this kind of thing, see in comparison to #Ashwini's accumu:
In [136]: timeit list(accumu(range(1000)))
10000 loops, best of 3: 161 us per loop
In [137]: timeit list(accumu(xrange(1000)))
10000 loops, best of 3: 147 us per loop
In [138]: timeit np.cumsum(np.arange(1000))
100000 loops, best of 3: 10.1 us per loop
But of course if it's the only place you'll use numpy, it might not be worth having a dependence on it.
In Python 2 you can define your own generator function like this:
def accumu(lis):
total = 0
for x in lis:
total += x
yield total
In [4]: list(accumu([4,6,12]))
Out[4]: [4, 10, 22]
And in Python 3.2+ you can use itertools.accumulate():
In [1]: lis = [4,6,12]
In [2]: from itertools import accumulate
In [3]: list(accumulate(lis))
Out[3]: [4, 10, 22]
I did a bench-mark of the top two answers with Python 3.4 and I found itertools.accumulate is faster than numpy.cumsum under many circumstances, often much faster. However, as you can see from the comments, this may not always be the case, and it's difficult to exhaustively explore all options. (Feel free to add a comment or edit this post if you have further benchmark results of interest.)
Some timings...
For short lists accumulate is about 4 times faster:
from timeit import timeit
def sum1(l):
from itertools import accumulate
return list(accumulate(l))
def sum2(l):
from numpy import cumsum
return list(cumsum(l))
l = [1, 2, 3, 4, 5]
timeit(lambda: sum1(l), number=100000)
# 0.4243644131347537
timeit(lambda: sum2(l), number=100000)
# 1.7077815784141421
For longer lists accumulate is about 3 times faster:
l = [1, 2, 3, 4, 5]*1000
timeit(lambda: sum1(l), number=100000)
# 19.174508565105498
timeit(lambda: sum2(l), number=100000)
# 61.871223849244416
If the numpy array is not cast to list, accumulate is still about 2 times faster:
from timeit import timeit
def sum1(l):
from itertools import accumulate
return list(accumulate(l))
def sum2(l):
from numpy import cumsum
return cumsum(l)
l = [1, 2, 3, 4, 5]*1000
print(timeit(lambda: sum1(l), number=100000))
# 19.18597290944308
print(timeit(lambda: sum2(l), number=100000))
# 37.759664884768426
If you put the imports outside of the two functions and still return a numpy array, accumulate is still nearly 2 times faster:
from timeit import timeit
from itertools import accumulate
from numpy import cumsum
def sum1(l):
return list(accumulate(l))
def sum2(l):
return cumsum(l)
l = [1, 2, 3, 4, 5]*1000
timeit(lambda: sum1(l), number=100000)
# 19.042188624851406
timeit(lambda: sum2(l), number=100000)
# 35.17324400227517
Try the
itertools.accumulate() function.
import itertools
list(itertools.accumulate([1,2,3,4,5]))
# [1, 3, 6, 10, 15]
Behold:
a = [4, 6, 12]
reduce(lambda c, x: c + [c[-1] + x], a, [0])[1:]
Will output (as expected):
[4, 10, 22]
Assignment expressions from PEP 572 (new in Python 3.8) offer yet another way to solve this:
time_interval = [4, 6, 12]
total_time = 0
cum_time = [total_time := total_time + t for t in time_interval]
You can calculate the cumulative sum list in linear time with a simple for loop:
def csum(lst):
s = lst.copy()
for i in range(1, len(s)):
s[i] += s[i-1]
return s
time_interval = [4, 6, 12]
print(csum(time_interval)) # [4, 10, 22]
The standard library's itertools.accumulate may be a faster alternative (since it's implemented in C):
from itertools import accumulate
time_interval = [4, 6, 12]
print(list(accumulate(time_interval))) # [4, 10, 22]
Since python 3.8 it's possible to use Assignment expressions, so things like this became easier to implement
nums = list(range(1, 10))
print(f'array: {nums}')
v = 0
cumsum = [v := v + n for n in nums]
print(f'cumsum: {cumsum}')
produces
array: [1, 2, 3, 4, 5, 6, 7, 8, 9]
cumsum: [1, 3, 6, 10, 15, 21, 28, 36, 45]
The same technique can be applied to find the cum product, mean, etc.
p = 1
cumprod = [p := p * n for n in nums]
print(f'cumprod: {cumprod}')
s = 0
c = 0
cumavg = [(s := s + n) / (c := c + 1) for n in nums]
print(f'cumavg: {cumavg}')
results in
cumprod: [1, 2, 6, 24, 120, 720, 5040, 40320, 362880]
cumavg: [1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0]
First, you want a running list of subsequences:
subseqs = (seq[:i] for i in range(1, len(seq)+1))
Then you just call sum on each subsequence:
sums = [sum(subseq) for subseq in subseqs]
(This isn't the most efficient way to do it, because you're adding all of the prefixes repeatedly. But that probably won't matter for most use cases, and it's easier to understand if you don't have to think of the running totals.)
If you're using Python 3.2 or newer, you can use itertools.accumulate to do it for you:
sums = itertools.accumulate(seq)
And if you're using 3.1 or earlier, you can just copy the "equivalent to" source straight out of the docs (except for changing next(it) to it.next() for 2.5 and earlier).
If You want a pythonic way without numpy working in 2.7 this would be my way of doing it
l = [1,2,3,4]
_d={-1:0}
cumsum=[_d.setdefault(idx, _d[idx-1]+item) for idx,item in enumerate(l)]
now let's try it and test it against all other implementations
import timeit, sys
L=list(range(10000))
if sys.version_info >= (3, 0):
reduce = functools.reduce
xrange = range
def sum1(l):
cumsum=[]
total = 0
for v in l:
total += v
cumsum.append(total)
return cumsum
def sum2(l):
import numpy as np
return list(np.cumsum(l))
def sum3(l):
return [sum(l[:i+1]) for i in xrange(len(l))]
def sum4(l):
return reduce(lambda c, x: c + [c[-1] + x], l, [0])[1:]
def this_implementation(l):
_d={-1:0}
return [_d.setdefault(idx, _d[idx-1]+item) for idx,item in enumerate(l)]
# sanity check
sum1(L)==sum2(L)==sum3(L)==sum4(L)==this_implementation(L)
>>> True
# PERFORMANCE TEST
timeit.timeit('sum1(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.001018061637878418
timeit.timeit('sum2(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.000829620361328125
timeit.timeit('sum3(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.4606760001182556
timeit.timeit('sum4(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.18932826995849608
timeit.timeit('this_implementation(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.002348129749298096
There could be many answers for this depending on the length of the list and the performance. One very simple way which I can think without thinking of the performance is this:
a = [1, 2, 3, 4]
a = [sum(a[0:x]) for x in range(1, len(a)+1)]
print(a)
[1, 3, 6, 10]
This is by using list comprehension and this may work fairly well it is just that here I am adding over the subarray many times, you could possibly improvise on this and make it simple!
Cheers to your endeavor!
values = [4, 6, 12]
total = 0
sums = []
for v in values:
total = total + v
sums.append(total)
print 'Values: ', values
print 'Sums: ', sums
Running this code gives
Values: [4, 6, 12]
Sums: [4, 10, 22]
Try this:
result = []
acc = 0
for i in time_interval:
acc += i
result.append(acc)
l = [1,-1,3]
cum_list = l
def sum_list(input_list):
index = 1
for i in input_list[1:]:
cum_list[index] = i + input_list[index-1]
index = index + 1
return cum_list
print(sum_list(l))
In Python3, To find the cumulative sum of a list where the ith element
is the sum of the first i+1 elements from the original list, you may do:
a = [4 , 6 , 12]
b = []
for i in range(0,len(a)):
b.append(sum(a[:i+1]))
print(b)
OR you may use list comprehension:
b = [sum(a[:x+1]) for x in range(0,len(a))]
Output
[4,10,22]
lst = [4, 6, 12]
[sum(lst[:i+1]) for i in xrange(len(lst))]
If you are looking for a more efficient solution (bigger lists?) a generator could be a good call (or just use numpy if you really care about performance).
def gen(lst):
acu = 0
for num in lst:
yield num + acu
acu += num
print list(gen([4, 6, 12]))
In [42]: a = [4, 6, 12]
In [43]: [sum(a[:i+1]) for i in xrange(len(a))]
Out[43]: [4, 10, 22]
This is slighlty faster than the generator method above by #Ashwini for small lists
In [48]: %timeit list(accumu([4,6,12]))
100000 loops, best of 3: 2.63 us per loop
In [49]: %timeit [sum(a[:i+1]) for i in xrange(len(a))]
100000 loops, best of 3: 2.46 us per loop
For larger lists, the generator is the way to go for sure. . .
In [50]: a = range(1000)
In [51]: %timeit [sum(a[:i+1]) for i in xrange(len(a))]
100 loops, best of 3: 6.04 ms per loop
In [52]: %timeit list(accumu(a))
10000 loops, best of 3: 162 us per loop
Somewhat hacky, but seems to work:
def cumulative_sum(l):
y = [0]
def inc(n):
y[0] += n
return y[0]
return [inc(x) for x in l]
I did think that the inner function would be able to modify the y declared in the outer lexical scope, but that didn't work, so we play some nasty hacks with structure modification instead. It is probably more elegant to use a generator.
Without having to use Numpy, you can loop directly over the array and accumulate the sum along the way. For example:
a=range(10)
i=1
while((i>0) & (i<10)):
a[i]=a[i-1]+a[i]
i=i+1
print a
Results in:
[0, 1, 3, 6, 10, 15, 21, 28, 36, 45]
A pure python oneliner for cumulative sum:
cumsum = lambda X: X[:1] + cumsum([X[0]+X[1]] + X[2:]) if X[1:] else X
This is a recursive version inspired by recursive cumulative sums. Some explanations:
The first term X[:1] is a list containing the previous element and is almost the same as [X[0]] (which would complain for empty lists).
The recursive cumsum call in the second term processes the current element [1] and remaining list whose length will be reduced by one.
if X[1:] is shorter for if len(X)>1.
Test:
cumsum([4,6,12])
#[4, 10, 22]
cumsum([])
#[]
And simular for cumulative product:
cumprod = lambda X: X[:1] + cumprod([X[0]*X[1]] + X[2:]) if X[1:] else X
Test:
cumprod([4,6,12])
#[4, 24, 288]
Here's another fun solution. This takes advantage of the locals() dict of a comprehension, i.e. local variables generated inside the list comprehension scope:
>>> [locals().setdefault(i, (elem + locals().get(i-1, 0))) for i, elem
in enumerate(time_interval)]
[4, 10, 22]
Here's what the locals() looks for each iteration:
>>> [[locals().setdefault(i, (elem + locals().get(i-1, 0))), locals().copy()][1]
for i, elem in enumerate(time_interval)]
[{'.0': <enumerate at 0x21f21f7fc80>, 'i': 0, 'elem': 4, 0: 4},
{'.0': <enumerate at 0x21f21f7fc80>, 'i': 1, 'elem': 6, 0: 4, 1: 10},
{'.0': <enumerate at 0x21f21f7fc80>, 'i': 2, 'elem': 12, 0: 4, 1: 10, 2: 22}]
Performance is not terrible for small lists:
>>> %timeit list(accumulate([4, 6, 12]))
387 ns ± 7.53 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)
>>> %timeit np.cumsum([4, 6, 12])
5.31 µs ± 67.8 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
>>> %timeit [locals().setdefault(i, (e + locals().get(i-1,0))) for i,e in enumerate(time_interval)]
1.57 µs ± 12 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)
And obviously falls flat for larger lists.
>>> l = list(range(1_000_000))
>>> %timeit list(accumulate(l))
95.1 ms ± 5.22 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
>>> %timeit np.cumsum(l)
79.3 ms ± 1.07 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
>>> %timeit np.cumsum(l).tolist()
120 ms ± 1.23 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
>>> %timeit [locals().setdefault(i, (e + locals().get(i-1, 0))) for i, e in enumerate(l)]
660 ms ± 5.14 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
Even though the method is ugly and not practical, it sure is fun.
I think the below code is the easiest:
a=[1,1,2,1,2]
b=[a[0]]+[sum(a[0:i]) for i in range(2,len(a)+1)]
def cumulative_sum(list):
l = []
for i in range(len(list)):
new_l = sum(list[:i+1])
l.append(new_l)
return l
time_interval = [4, 6, 12]
print(cumulative_sum(time_interval)
Maybe a more beginner-friendly solution.
So you need to make a list of cumulative sums. You can do it by using for loop and .append() method
time_interval = [4, 6, 12]
cumulative_sum = []
new_sum = 0
for i in time_interval:
new_sum += i
cumulative_sum.append(new_sum)
print(cumulative_sum)
or, using numpy module
import numpy
time_interval = [4, 6, 12]
c_sum = numpy.cumsum(time_interval)
print(c_sum.tolist())
This would be Haskell-style:
def wrand(vtlg):
def helpf(lalt,lneu):
if not lalt==[]:
return helpf(lalt[1::],[lalt[0]+lneu[0]]+lneu)
else:
lneu.reverse()
return lneu[1:]
return helpf(vtlg,[0])
time_interval = [4, 6, 12]
I want to sum up the numbers like [4, 4+6, 4+6+12] in order to get the list t = [4, 10, 22].
I tried the following:
t1 = time_interval[0]
t2 = time_interval[1] + t1
t3 = time_interval[2] + t2
print(t1, t2, t3) # -> 4 10 22
If you're doing much numerical work with arrays like this, I'd suggest numpy, which comes with a cumulative sum function cumsum:
import numpy as np
a = [4,6,12]
np.cumsum(a)
#array([4, 10, 22])
Numpy is often faster than pure python for this kind of thing, see in comparison to #Ashwini's accumu:
In [136]: timeit list(accumu(range(1000)))
10000 loops, best of 3: 161 us per loop
In [137]: timeit list(accumu(xrange(1000)))
10000 loops, best of 3: 147 us per loop
In [138]: timeit np.cumsum(np.arange(1000))
100000 loops, best of 3: 10.1 us per loop
But of course if it's the only place you'll use numpy, it might not be worth having a dependence on it.
In Python 2 you can define your own generator function like this:
def accumu(lis):
total = 0
for x in lis:
total += x
yield total
In [4]: list(accumu([4,6,12]))
Out[4]: [4, 10, 22]
And in Python 3.2+ you can use itertools.accumulate():
In [1]: lis = [4,6,12]
In [2]: from itertools import accumulate
In [3]: list(accumulate(lis))
Out[3]: [4, 10, 22]
I did a bench-mark of the top two answers with Python 3.4 and I found itertools.accumulate is faster than numpy.cumsum under many circumstances, often much faster. However, as you can see from the comments, this may not always be the case, and it's difficult to exhaustively explore all options. (Feel free to add a comment or edit this post if you have further benchmark results of interest.)
Some timings...
For short lists accumulate is about 4 times faster:
from timeit import timeit
def sum1(l):
from itertools import accumulate
return list(accumulate(l))
def sum2(l):
from numpy import cumsum
return list(cumsum(l))
l = [1, 2, 3, 4, 5]
timeit(lambda: sum1(l), number=100000)
# 0.4243644131347537
timeit(lambda: sum2(l), number=100000)
# 1.7077815784141421
For longer lists accumulate is about 3 times faster:
l = [1, 2, 3, 4, 5]*1000
timeit(lambda: sum1(l), number=100000)
# 19.174508565105498
timeit(lambda: sum2(l), number=100000)
# 61.871223849244416
If the numpy array is not cast to list, accumulate is still about 2 times faster:
from timeit import timeit
def sum1(l):
from itertools import accumulate
return list(accumulate(l))
def sum2(l):
from numpy import cumsum
return cumsum(l)
l = [1, 2, 3, 4, 5]*1000
print(timeit(lambda: sum1(l), number=100000))
# 19.18597290944308
print(timeit(lambda: sum2(l), number=100000))
# 37.759664884768426
If you put the imports outside of the two functions and still return a numpy array, accumulate is still nearly 2 times faster:
from timeit import timeit
from itertools import accumulate
from numpy import cumsum
def sum1(l):
return list(accumulate(l))
def sum2(l):
return cumsum(l)
l = [1, 2, 3, 4, 5]*1000
timeit(lambda: sum1(l), number=100000)
# 19.042188624851406
timeit(lambda: sum2(l), number=100000)
# 35.17324400227517
Try the
itertools.accumulate() function.
import itertools
list(itertools.accumulate([1,2,3,4,5]))
# [1, 3, 6, 10, 15]
Behold:
a = [4, 6, 12]
reduce(lambda c, x: c + [c[-1] + x], a, [0])[1:]
Will output (as expected):
[4, 10, 22]
Assignment expressions from PEP 572 (new in Python 3.8) offer yet another way to solve this:
time_interval = [4, 6, 12]
total_time = 0
cum_time = [total_time := total_time + t for t in time_interval]
You can calculate the cumulative sum list in linear time with a simple for loop:
def csum(lst):
s = lst.copy()
for i in range(1, len(s)):
s[i] += s[i-1]
return s
time_interval = [4, 6, 12]
print(csum(time_interval)) # [4, 10, 22]
The standard library's itertools.accumulate may be a faster alternative (since it's implemented in C):
from itertools import accumulate
time_interval = [4, 6, 12]
print(list(accumulate(time_interval))) # [4, 10, 22]
Since python 3.8 it's possible to use Assignment expressions, so things like this became easier to implement
nums = list(range(1, 10))
print(f'array: {nums}')
v = 0
cumsum = [v := v + n for n in nums]
print(f'cumsum: {cumsum}')
produces
array: [1, 2, 3, 4, 5, 6, 7, 8, 9]
cumsum: [1, 3, 6, 10, 15, 21, 28, 36, 45]
The same technique can be applied to find the cum product, mean, etc.
p = 1
cumprod = [p := p * n for n in nums]
print(f'cumprod: {cumprod}')
s = 0
c = 0
cumavg = [(s := s + n) / (c := c + 1) for n in nums]
print(f'cumavg: {cumavg}')
results in
cumprod: [1, 2, 6, 24, 120, 720, 5040, 40320, 362880]
cumavg: [1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0]
First, you want a running list of subsequences:
subseqs = (seq[:i] for i in range(1, len(seq)+1))
Then you just call sum on each subsequence:
sums = [sum(subseq) for subseq in subseqs]
(This isn't the most efficient way to do it, because you're adding all of the prefixes repeatedly. But that probably won't matter for most use cases, and it's easier to understand if you don't have to think of the running totals.)
If you're using Python 3.2 or newer, you can use itertools.accumulate to do it for you:
sums = itertools.accumulate(seq)
And if you're using 3.1 or earlier, you can just copy the "equivalent to" source straight out of the docs (except for changing next(it) to it.next() for 2.5 and earlier).
If You want a pythonic way without numpy working in 2.7 this would be my way of doing it
l = [1,2,3,4]
_d={-1:0}
cumsum=[_d.setdefault(idx, _d[idx-1]+item) for idx,item in enumerate(l)]
now let's try it and test it against all other implementations
import timeit, sys
L=list(range(10000))
if sys.version_info >= (3, 0):
reduce = functools.reduce
xrange = range
def sum1(l):
cumsum=[]
total = 0
for v in l:
total += v
cumsum.append(total)
return cumsum
def sum2(l):
import numpy as np
return list(np.cumsum(l))
def sum3(l):
return [sum(l[:i+1]) for i in xrange(len(l))]
def sum4(l):
return reduce(lambda c, x: c + [c[-1] + x], l, [0])[1:]
def this_implementation(l):
_d={-1:0}
return [_d.setdefault(idx, _d[idx-1]+item) for idx,item in enumerate(l)]
# sanity check
sum1(L)==sum2(L)==sum3(L)==sum4(L)==this_implementation(L)
>>> True
# PERFORMANCE TEST
timeit.timeit('sum1(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.001018061637878418
timeit.timeit('sum2(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.000829620361328125
timeit.timeit('sum3(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.4606760001182556
timeit.timeit('sum4(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.18932826995849608
timeit.timeit('this_implementation(L)','from __main__ import sum1,sum2,sum3,sum4,this_implementation,L', number=100)/100.
>>> 0.002348129749298096
There could be many answers for this depending on the length of the list and the performance. One very simple way which I can think without thinking of the performance is this:
a = [1, 2, 3, 4]
a = [sum(a[0:x]) for x in range(1, len(a)+1)]
print(a)
[1, 3, 6, 10]
This is by using list comprehension and this may work fairly well it is just that here I am adding over the subarray many times, you could possibly improvise on this and make it simple!
Cheers to your endeavor!
values = [4, 6, 12]
total = 0
sums = []
for v in values:
total = total + v
sums.append(total)
print 'Values: ', values
print 'Sums: ', sums
Running this code gives
Values: [4, 6, 12]
Sums: [4, 10, 22]
Try this:
result = []
acc = 0
for i in time_interval:
acc += i
result.append(acc)
l = [1,-1,3]
cum_list = l
def sum_list(input_list):
index = 1
for i in input_list[1:]:
cum_list[index] = i + input_list[index-1]
index = index + 1
return cum_list
print(sum_list(l))
In Python3, To find the cumulative sum of a list where the ith element
is the sum of the first i+1 elements from the original list, you may do:
a = [4 , 6 , 12]
b = []
for i in range(0,len(a)):
b.append(sum(a[:i+1]))
print(b)
OR you may use list comprehension:
b = [sum(a[:x+1]) for x in range(0,len(a))]
Output
[4,10,22]
lst = [4, 6, 12]
[sum(lst[:i+1]) for i in xrange(len(lst))]
If you are looking for a more efficient solution (bigger lists?) a generator could be a good call (or just use numpy if you really care about performance).
def gen(lst):
acu = 0
for num in lst:
yield num + acu
acu += num
print list(gen([4, 6, 12]))
In [42]: a = [4, 6, 12]
In [43]: [sum(a[:i+1]) for i in xrange(len(a))]
Out[43]: [4, 10, 22]
This is slighlty faster than the generator method above by #Ashwini for small lists
In [48]: %timeit list(accumu([4,6,12]))
100000 loops, best of 3: 2.63 us per loop
In [49]: %timeit [sum(a[:i+1]) for i in xrange(len(a))]
100000 loops, best of 3: 2.46 us per loop
For larger lists, the generator is the way to go for sure. . .
In [50]: a = range(1000)
In [51]: %timeit [sum(a[:i+1]) for i in xrange(len(a))]
100 loops, best of 3: 6.04 ms per loop
In [52]: %timeit list(accumu(a))
10000 loops, best of 3: 162 us per loop
Somewhat hacky, but seems to work:
def cumulative_sum(l):
y = [0]
def inc(n):
y[0] += n
return y[0]
return [inc(x) for x in l]
I did think that the inner function would be able to modify the y declared in the outer lexical scope, but that didn't work, so we play some nasty hacks with structure modification instead. It is probably more elegant to use a generator.
Without having to use Numpy, you can loop directly over the array and accumulate the sum along the way. For example:
a=range(10)
i=1
while((i>0) & (i<10)):
a[i]=a[i-1]+a[i]
i=i+1
print a
Results in:
[0, 1, 3, 6, 10, 15, 21, 28, 36, 45]
A pure python oneliner for cumulative sum:
cumsum = lambda X: X[:1] + cumsum([X[0]+X[1]] + X[2:]) if X[1:] else X
This is a recursive version inspired by recursive cumulative sums. Some explanations:
The first term X[:1] is a list containing the previous element and is almost the same as [X[0]] (which would complain for empty lists).
The recursive cumsum call in the second term processes the current element [1] and remaining list whose length will be reduced by one.
if X[1:] is shorter for if len(X)>1.
Test:
cumsum([4,6,12])
#[4, 10, 22]
cumsum([])
#[]
And simular for cumulative product:
cumprod = lambda X: X[:1] + cumprod([X[0]*X[1]] + X[2:]) if X[1:] else X
Test:
cumprod([4,6,12])
#[4, 24, 288]
Here's another fun solution. This takes advantage of the locals() dict of a comprehension, i.e. local variables generated inside the list comprehension scope:
>>> [locals().setdefault(i, (elem + locals().get(i-1, 0))) for i, elem
in enumerate(time_interval)]
[4, 10, 22]
Here's what the locals() looks for each iteration:
>>> [[locals().setdefault(i, (elem + locals().get(i-1, 0))), locals().copy()][1]
for i, elem in enumerate(time_interval)]
[{'.0': <enumerate at 0x21f21f7fc80>, 'i': 0, 'elem': 4, 0: 4},
{'.0': <enumerate at 0x21f21f7fc80>, 'i': 1, 'elem': 6, 0: 4, 1: 10},
{'.0': <enumerate at 0x21f21f7fc80>, 'i': 2, 'elem': 12, 0: 4, 1: 10, 2: 22}]
Performance is not terrible for small lists:
>>> %timeit list(accumulate([4, 6, 12]))
387 ns ± 7.53 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)
>>> %timeit np.cumsum([4, 6, 12])
5.31 µs ± 67.8 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
>>> %timeit [locals().setdefault(i, (e + locals().get(i-1,0))) for i,e in enumerate(time_interval)]
1.57 µs ± 12 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)
And obviously falls flat for larger lists.
>>> l = list(range(1_000_000))
>>> %timeit list(accumulate(l))
95.1 ms ± 5.22 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
>>> %timeit np.cumsum(l)
79.3 ms ± 1.07 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
>>> %timeit np.cumsum(l).tolist()
120 ms ± 1.23 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
>>> %timeit [locals().setdefault(i, (e + locals().get(i-1, 0))) for i, e in enumerate(l)]
660 ms ± 5.14 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
Even though the method is ugly and not practical, it sure is fun.
I think the below code is the easiest:
a=[1,1,2,1,2]
b=[a[0]]+[sum(a[0:i]) for i in range(2,len(a)+1)]
def cumulative_sum(list):
l = []
for i in range(len(list)):
new_l = sum(list[:i+1])
l.append(new_l)
return l
time_interval = [4, 6, 12]
print(cumulative_sum(time_interval)
Maybe a more beginner-friendly solution.
So you need to make a list of cumulative sums. You can do it by using for loop and .append() method
time_interval = [4, 6, 12]
cumulative_sum = []
new_sum = 0
for i in time_interval:
new_sum += i
cumulative_sum.append(new_sum)
print(cumulative_sum)
or, using numpy module
import numpy
time_interval = [4, 6, 12]
c_sum = numpy.cumsum(time_interval)
print(c_sum.tolist())
This would be Haskell-style:
def wrand(vtlg):
def helpf(lalt,lneu):
if not lalt==[]:
return helpf(lalt[1::],[lalt[0]+lneu[0]]+lneu)
else:
lneu.reverse()
return lneu[1:]
return helpf(vtlg,[0])
Does range function allows concatenation ? Like i want to make a range(30) & concatenate it with range(2000, 5002). So my concatenated range will be 0, 1, 2, ... 29, 2000, 2001, ... 5001
Code like this does not work on my latest python (ver: 3.3.0)
range(30) + range(2000, 5002)
You can use itertools.chain for this:
from itertools import chain
concatenated = chain(range(30), range(2000, 5002))
for i in concatenated:
...
It works for arbitrary iterables. Note that there's a difference in behavior of range() between Python 2 and 3 that you should know about: in Python 2 range returns a list, and in Python3 an iterator, which is memory-efficient, but not always desirable.
Lists can be concatenated with +, iterators cannot.
I like the most simple solutions that are possible (including efficiency). It is not always clear whether the solution is such. Anyway, the range() in Python 3 is a generator. You can wrap it to any construct that does iteration. The list() is capable of construction of a list value from any iterable. The + operator for lists does concatenation. I am using smaller values in the example:
>>> list(range(5))
[0, 1, 2, 3, 4]
>>> list(range(10, 20))
[10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
>>> list(range(5)) + list(range(10,20))
[0, 1, 2, 3, 4, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
This is what range(5) + range(10, 20) exactly did in Python 2.5 -- because range() returned a list.
In Python 3, it is only useful if you really want to construct the list. Otherwise, I recommend the Lev Levitsky's solution with itertools.chain. The documentation also shows the very straightforward implementation:
def chain(*iterables):
# chain('ABC', 'DEF') --> A B C D E F
for it in iterables:
for element in it:
yield element
The solution by Inbar Rose is fine and functionally equivalent. Anyway, my +1 goes to Lev Levitsky and to his argument about using the standard libraries. From The Zen of Python...
In the face of ambiguity, refuse the temptation to guess.
#!python3
import timeit
number = 10000
t = timeit.timeit('''\
for i in itertools.chain(range(30), range(2000, 5002)):
pass
''',
'import itertools', number=number)
print('itertools:', t/number * 1000000, 'microsec/one execution')
t = timeit.timeit('''\
for x in (i for j in (range(30), range(2000, 5002)) for i in j):
pass
''', number=number)
print('generator expression:', t/number * 1000000, 'microsec/one execution')
In my opinion, the itertools.chain is more readable. But what really is important...
itertools: 264.4522138986938 microsec/one execution
generator expression: 785.3081048010291 microsec/one execution
... it is about 3 times faster.
python >= 3.5
You can use iterable unpacking in lists (see PEP 448: Additional Unpacking Generalizations).
If you need a list,
[*range(2, 5), *range(3, 7)]
# [2, 3, 4, 3, 4, 5, 6]
This preserves order and does not remove duplicates. Or, you might want a tuple,
(*range(2, 5), *range(3, 7))
# (2, 3, 4, 3, 4, 5, 6)
... or a set,
# note that this drops duplicates
{*range(2, 5), *range(3, 7)}
# {2, 3, 4, 5, 6}
It also happens to be faster than calling itertools.chain.
from itertools import chain
%timeit list(chain(range(10000), range(5000, 20000)))
%timeit [*range(10000), *range(5000, 20000)]
738 µs ± 10.4 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
665 µs ± 13.8 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
The benefit of chain, however, is that you can pass an arbitrary list of ranges.
ranges = [range(2, 5), range(3, 7), ...]
flat = list(chain.from_iterable(ranges))
OTOH, unpacking generalisations haven't been "generalised" to arbitrary sequences, so you will still need to unpack the individual ranges yourself.
Can be done using list-comprehension.
>>> [i for j in (range(10), range(15, 20)) for i in j]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 15, 16, 17, 18, 19]
Works for your request, but it is a long answer so I will not post it here.
note: can be made into a generator for increased performance:
for x in (i for j in (range(30), range(2000, 5002)) for i in j):
# code
or even into a generator variable.
gen = (i for j in (range(30), range(2000, 5002)) for i in j)
for x in gen:
# code
With the help of the extend method, we can concatenate two lists.
>>> a = list(range(1,10))
>>> a.extend(range(100,105))
>>> a
[1, 2, 3, 4, 5, 6, 7, 8, 9, 100, 101, 102, 103, 104]
range() in Python 2.x returns a list:
>>> range(10)
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
xrange() in Python 2.x returns an iterator:
>>> xrange(10)
xrange(10)
And in Python 3 range() also returns an iterator:
>>> r = range(10)
>>> iterator = r.__iter__()
>>> iterator.__next__()
0
>>> iterator.__next__()
1
>>> iterator.__next__()
2
So it is clear that you can not concatenate iterators other by using chain() as the other guy pointed out.
You can use list function around range function to make a list
LIKE THIS
list(range(3,7))+list(range(2,9))
I came to this question because I was trying to concatenate an unknown number of ranges, that might overlap, and didn't want repeated values in the final iterator. My solution was to use set and the union operator like so:
range1 = range(1,4)
range2 = range(2,6)
concatenated = set.union(set(range1), set(range2)
for i in concatenated:
print(i)