I can't figure out how to look ahead one element in a Python generator. As soon as I look it's gone.
Here is what I mean:
gen = iter([1,2,3])
next_value = gen.next() # okay, I looked forward and see that next_value = 1
# but now:
list(gen) # is [2, 3] -- the first value is gone!
Here is a more real example:
gen = element_generator()
if gen.next_value() == 'STOP':
quit_application()
else:
process(gen.next())
Can anyone help me write a generator that you can look one element forward?
See also: Resetting generator object in Python
For sake of completeness, the more-itertools package (which should probably be part of any Python programmer's toolbox) includes a peekable wrapper that implements this behavior. As the code example in the documentation shows:
>>> p = peekable(['a', 'b'])
>>> p.peek()
'a'
>>> next(p)
'a'
However, it's often possible to rewrite code that would use this functionality so that it doesn't actually need it. For example, your realistic code sample from the question could be written like this:
gen = element_generator()
command = gen.next_value()
if command == 'STOP':
quit_application()
else:
process(command)
(reader's note: I've preserved the syntax in the example from the question as of when I'm writing this, even though it refers to an outdated version of Python)
The Python generator API is one way: You can't push back elements you've read. But you can create a new iterator using the itertools module and prepend the element:
import itertools
gen = iter([1,2,3])
peek = gen.next()
print list(itertools.chain([peek], gen))
Ok - two years too late - but I came across this question, and did not find any of the answers to my satisfaction. Came up with this meta generator:
class Peekorator(object):
def __init__(self, generator):
self.empty = False
self.peek = None
self.generator = generator
try:
self.peek = self.generator.next()
except StopIteration:
self.empty = True
def __iter__(self):
return self
def next(self):
"""
Return the self.peek element, or raise StopIteration
if empty
"""
if self.empty:
raise StopIteration()
to_return = self.peek
try:
self.peek = self.generator.next()
except StopIteration:
self.peek = None
self.empty = True
return to_return
def simple_iterator():
for x in range(10):
yield x*3
pkr = Peekorator(simple_iterator())
for i in pkr:
print i, pkr.peek, pkr.empty
results in:
0 3 False
3 6 False
6 9 False
9 12 False
...
24 27 False
27 None False
i.e. you have at any moment during iteration access to the next item in the list.
Using itertools.tee will produce a lightweight copy of the generator; then peeking ahead at one copy will not affect the second copy. Thus:
import itertools
def process(seq):
peeker, items = itertools.tee(seq)
# initial peek ahead
# so that peeker is one ahead of items
if next(peeker) == 'STOP':
return
for item in items:
# peek ahead
if next(peeker) == "STOP":
return
# process items
print(item)
The items generator is unaffected by modifications to peeker. However, modifying seq after the call to tee may cause problems.
That said: any algorithm that requires looking an item ahead in a generator could instead be written to use the current generator item and the previous item. This will result in simpler code - see my other answer to this question.
An iterator that allows peeking at the next element and also further ahead. It reads ahead as needed and remembers the values in a deque.
from collections import deque
class PeekIterator:
def __init__(self, iterable):
self.iterator = iter(iterable)
self.peeked = deque()
def __iter__(self):
return self
def __next__(self):
if self.peeked:
return self.peeked.popleft()
return next(self.iterator)
def peek(self, ahead=0):
while len(self.peeked) <= ahead:
self.peeked.append(next(self.iterator))
return self.peeked[ahead]
Demo:
>>> it = PeekIterator(range(10))
>>> it.peek()
0
>>> it.peek(5)
5
>>> it.peek(13)
Traceback (most recent call last):
File "<pyshell#68>", line 1, in <module>
it.peek(13)
File "[...]", line 15, in peek
self.peeked.append(next(self.iterator))
StopIteration
>>> it.peek(2)
2
>>> next(it)
0
>>> it.peek(2)
3
>>> list(it)
[1, 2, 3, 4, 5, 6, 7, 8, 9]
>>>
>>> gen = iter(range(10))
>>> peek = next(gen)
>>> peek
0
>>> gen = (value for g in ([peek], gen) for value in g)
>>> list(gen)
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
Just for fun, I created an implementation of a lookahead class based on the suggestion by
Aaron:
import itertools
class lookahead_chain(object):
def __init__(self, it):
self._it = iter(it)
def __iter__(self):
return self
def next(self):
return next(self._it)
def peek(self, default=None, _chain=itertools.chain):
it = self._it
try:
v = self._it.next()
self._it = _chain((v,), it)
return v
except StopIteration:
return default
lookahead = lookahead_chain
With this, the following will work:
>>> t = lookahead(xrange(8))
>>> list(itertools.islice(t, 3))
[0, 1, 2]
>>> t.peek()
3
>>> list(itertools.islice(t, 3))
[3, 4, 5]
With this implementation it is a bad idea to call peek many times in a row...
While looking at the CPython source code I just found a better way which is both shorter and more efficient:
class lookahead_tee(object):
def __init__(self, it):
self._it, = itertools.tee(it, 1)
def __iter__(self):
return self._it
def peek(self, default=None):
try:
return self._it.__copy__().next()
except StopIteration:
return default
lookahead = lookahead_tee
Usage is the same as above but you won't pay a price here to use peek many times in a row. With a few more lines you can also look ahead more than one item in the iterator (up to available RAM).
A simple solution is to use a function like this:
def peek(it):
first = next(it)
return first, itertools.chain([first], it)
Then you can do:
>>> it = iter(range(10))
>>> x, it = peek(it)
>>> x
0
>>> next(it)
0
>>> next(it)
1
If anybody is interested, and please correct me if I am wrong, but I believe it is pretty easy to add some push back functionality to any iterator.
class Back_pushable_iterator:
"""Class whose constructor takes an iterator as its only parameter, and
returns an iterator that behaves in the same way, with added push back
functionality.
The idea is to be able to push back elements that need to be retrieved once
more with the iterator semantics. This is particularly useful to implement
LL(k) parsers that need k tokens of lookahead. Lookahead or push back is
really a matter of perspective. The pushing back strategy allows a clean
parser implementation based on recursive parser functions.
The invoker of this class takes care of storing the elements that should be
pushed back. A consequence of this is that any elements can be "pushed
back", even elements that have never been retrieved from the iterator.
The elements that are pushed back are then retrieved through the iterator
interface in a LIFO-manner (as should logically be expected).
This class works for any iterator but is especially meaningful for a
generator iterator, which offers no obvious push back ability.
In the LL(k) case mentioned above, the tokenizer can be implemented by a
standard generator function (clean and simple), that is completed by this
class for the needs of the actual parser.
"""
def __init__(self, iterator):
self.iterator = iterator
self.pushed_back = []
def __iter__(self):
return self
def __next__(self):
if self.pushed_back:
return self.pushed_back.pop()
else:
return next(self.iterator)
def push_back(self, element):
self.pushed_back.append(element)
it = Back_pushable_iterator(x for x in range(10))
x = next(it) # 0
print(x)
it.push_back(x)
x = next(it) # 0
print(x)
x = next(it) # 1
print(x)
x = next(it) # 2
y = next(it) # 3
print(x)
print(y)
it.push_back(y)
it.push_back(x)
x = next(it) # 2
y = next(it) # 3
print(x)
print(y)
for x in it:
print(x) # 4-9
This will work -- it buffers an item and calls a function with each item and the next item in the sequence.
Your requirements are murky on what happens at the end of the sequence. What does "look ahead" mean when you're at the last one?
def process_with_lookahead( iterable, aFunction ):
prev= iterable.next()
for item in iterable:
aFunction( prev, item )
prev= item
aFunction( item, None )
def someLookaheadFunction( item, next_item ):
print item, next_item
Instead of using items (i, i+1), where 'i' is the current item and i+1 is the 'peek ahead' version, you should be using (i-1, i), where 'i-1' is the previous version from the generator.
Tweaking your algorithm this way will produce something that is identical to what you currently have, apart from the extra needless complexity of trying to 'peek ahead'.
Peeking ahead is a mistake, and you should not be doing it.
Although itertools.chain() is the natural tool for the job here, beware of loops like this:
for elem in gen:
...
peek = next(gen)
gen = itertools.chain([peek], gen)
...Because this will consume a linearly growing amount of memory, and eventually grind to a halt. (This code essentially seems to create a linked list, one node per chain() call.) I know this not because I inspected the libs but because this just resulted in a major slowdown of my program - getting rid of the gen = itertools.chain([peek], gen) line sped it up again. (Python 3.3)
Python3 snippet for #jonathan-hartley answer:
def peek(iterator, eoi=None):
iterator = iter(iterator)
try:
prev = next(iterator)
except StopIteration:
return iterator
for elm in iterator:
yield prev, elm
prev = elm
yield prev, eoi
for curr, nxt in peek(range(10)):
print((curr, nxt))
# (0, 1)
# (1, 2)
# (2, 3)
# (3, 4)
# (4, 5)
# (5, 6)
# (6, 7)
# (7, 8)
# (8, 9)
# (9, None)
It'd be straightforward to create a class that does this on __iter__ and yields just the prev item and put the elm in some attribute.
w.r.t #David Z's post, the newer seekable tool can reset a wrapped iterator to a prior position.
>>> s = mit.seekable(range(3))
>>> s.next()
# 0
>>> s.seek(0) # reset iterator
>>> s.next()
# 0
>>> s.next()
# 1
>>> s.seek(1)
>>> s.next()
# 1
>>> next(s)
# 2
cytoolz has a peek function.
>> from cytoolz import peek
>> gen = iter([1,2,3])
>> first, continuation = peek(gen)
>> first
1
>> list(continuation)
[1, 2, 3]
In my case, I need a generator where I could queue back to generator the data I have just got via next() call.
The way I handle this problem, is to create a queue. In the implementation of the generator, I would first check the queue: if queue is not empty, the "yield" will return the values in queue, or otherwise the values in normal way.
import queue
def gen1(n, q):
i = 0
while True:
if not q.empty():
yield q.get()
else:
yield i
i = i + 1
if i >= n:
if not q.empty():
yield q.get()
break
q = queue.Queue()
f = gen1(2, q)
i = next(f)
print(i)
i = next(f)
print(i)
q.put(i) # put back the value I have just got for following 'next' call
i = next(f)
print(i)
running
python3 gen_test.py
0
1
1
This concept is very useful when I was writing a parser, which needs to look the file line by line, if the line appears to belong to next phase of parsing, I could just queue back to the generator so that the next phase of code could parse it correctly without handling complex state.
For those of you who embrace frugality and one-liners, I present to you a one-liner that allows one to look ahead in an iterable (this only works in Python 3.8 and above):
>>> import itertools as it
>>> peek = lambda iterable, n=1: it.islice(zip(it.chain((t := it.tee(iterable))[0], [None] * n), it.chain([None] * n, t[1])), n, None)
>>> for lookahead, element in peek(range(10)):
... print(lookahead, element)
1 0
2 1
3 2
4 3
5 4
6 5
7 6
8 7
9 8
None 9
>>> for lookahead, element in peek(range(10), 2):
... print(lookahead, element)
2 0
3 1
4 2
5 3
6 4
7 5
8 6
9 7
None 8
None 9
This method is space-efficient by avoiding copying the iterator multiple times. It is also fast due to how it lazily generates elements. Finally, as a cherry on top, you can look ahead an arbitrary number of elements.
An algorithm that works by "peeking" at the next element in a generator could equivalently be one that works by remembering the previous element, treating that element as the one to operate upon, and treating the "current" element as simply "peeked at".
Either way, what is really happening is that the algorithm considers overlapping pairs from the generator. The itertools.tee recipe will work fine - and it is not hard to see that it is essentially a refactored version of Jonathan Hartley's approach:
from itertools import tee
# From https://docs.python.org/3/library/itertools.html#itertools.pairwise
# In 3.10 and up, this is directly supplied by the `itertools` module.
def pairwise(iterable):
# pairwise('ABCDEFG') --> AB BC CD DE EF FG
a, b = tee(iterable)
next(b, None)
return zip(a, b)
def process(seq):
for to_process, lookahead in pairwise(seq):
# peek ahead
if lookahead == "STOP":
return
# process items
print(to_process)
Using iter in our for loop makes programming an efficient one in python. How does it actually works?
Tried to visualize iter(iterables) in "http://www.pythontutor.com/visualize.html#mode=display". Here iter helps to create an instance.
Doesn't it actually refer to internal numerical objects?
val = [1,2,3,4,5]
val = iter(val)
for item in val:
print(item)
val = [1,2,3,4,5]
for item in val:
print(item)
Both returns same output. But how iter identifies the values?
What you're doing is redundant. A for loop is essentially syntactic sugar for:
val = [1,2,3,4,5]
iterator = iter(val)
while True:
try:
item = next(iterator)
except StopIteration:
break
print(item)
All that you're doing by calling iter() on your val is replacing
iterator = iter(val)
with
iterator = iter(iter(val))
Calling iter() on an iterator is a no-op; returning the same iterator.
Lets pretend that Python lists were not iterable and we wanted to create a class, MyList, that could be constructed with a list instance and iterate it. MyList would need to implement a __iter__ method that would return an iterator object that implements the __next__ method. Each successive call to this __next__ method should return the next element of the list. When there are no more elements to be returned, it needs to raise a StopIteration exception. This iterator object clearly needs two pieces of information:
A reference to the list being iterated.
The index of the last element that was output.
Let's call this iterator class MyListIterator. Its implementation might be:
class MyListIterator:
def __init__(self, l):
self.l = l # the list being iterated
self.index = -1 # the index of the last element outputted
def __next__(self):
self.index += 1 # next index to output
if self.index < len(self.l):
return self.l[self.index]
raise StopIteration
Class MyList then would be:
class MyList:
def __init__(self, the_list):
self.l = the_list # we are constructed with a list
def __iter__(self):
return MyListIterator(self.l) # pass to the iterator our list
An example of use:
l = MyList([0, 1, 2])
for i in l:
for j in l:
print (i, j)
Prints:
0 0
0 1
0 2
1 0
1 1
1 2
2 0
2 1
2 2
I would like to compare all elements in my iterable object combinatorically with each other. The following reproducible example just mimics the functionality of a plain list, but demonstrates my problem. In this example with a list of ["A","B","C","D"], I would like to get the following 16 lines of output, every combination of each item with each other. A list of 100 items should generate 100*100=10,000 lines.
A A True
A B False
A C False
... 10 more lines ...
D B False
D C False
D D True
The following code seemed like it should do the job.
class C():
def __init__(self):
self.stuff = ["A","B","C","D"]
def __iter__(self):
self.idx = 0
return self
def __next__(self):
self.idx += 1
if self.idx > len(self.stuff):
raise StopIteration
else:
return self.stuff[self.idx - 1]
thing = C()
for x in thing:
for y in thing:
print(x, y, x==y)
But after finishing the y-loop, the x-loop seems done, too, even though it's only used the first item in the iterable.
A A True
A B False
A C False
A D False
After much searching, I eventually tried the following code, hoping that itertools.tee would allow me two independent iterators over the same data:
import itertools
thing = C()
thing_one, thing_two = itertools.tee(thing)
for x in thing_one:
for y in thing_two:
print(x, y, x==y)
But I got the same output as before.
The real-world object this represents is a model of a directory and file structure with varying numbers of files and subdirectories, at varying depths into the tree. It has nested links to thousands of members and iterates correctly over them once, just like this example. But it also does expensive processing within its many internal objects on-the-fly as needed for comparisons, which would end up doubling the workload if I had to make a complete copy of it prior to iterating. I would really like to use multiple iterators, pointing into a single object with all the data, if possible.
Edit on answers: The critical flaw in the question code, pointed out in all answers, is the single internal self.idx variable being unable to handle multiple callers independently. The accepted answer is the best for my real class (oversimplified in this reproducible example), another answer presents a simple, elegant solution for simpler data structures like the list presented here.
It's actually impossible to make a container class that is it's own iterator. The container shouldn't know about the state of the iterator and the iterator doesn't need to know the contents of the container, it just needs to know which object is the corresponding container and "where" it is. If you mix iterator and container different iterators will share state with each other (in your case the self.idx) which will not give the correct results (they read and modify the same variable).
That's the reason why all built-in types have a seperate iterator class (and even some have an reverse-iterator class):
>>> l = [1, 2, 3]
>>> iter(l)
<list_iterator at 0x15e360c86d8>
>>> reversed(l)
<list_reverseiterator at 0x15e360a5940>
>>> t = (1, 2, 3)
>>> iter(t)
<tuple_iterator at 0x15e363fb320>
>>> s = '123'
>>> iter(s)
<str_iterator at 0x15e363fb438>
So, basically you could just return iter(self.stuff) in __iter__ and drop the __next__ altogether because list_iterator knows how to iterate over the list:
class C:
def __init__(self):
self.stuff = ["A","B","C","D"]
def __iter__(self):
return iter(self.stuff)
thing = C()
for x in thing:
for y in thing:
print(x, y, x==y)
prints 16 lines, like expected.
If your goal is to make your own iterator class, you need two classes (or 3 if you want to implement the reversed-iterator yourself).
class C:
def __init__(self):
self.stuff = ["A","B","C","D"]
def __iter__(self):
return C_iterator(self)
def __reversed__(self):
return C_reversed_iterator(self)
class C_iterator:
def __init__(self, parent):
self.idx = 0
self.parent = parent
def __iter__(self):
return self
def __next__(self):
self.idx += 1
if self.idx > len(self.parent.stuff):
raise StopIteration
else:
return self.parent.stuff[self.idx - 1]
thing = C()
for x in thing:
for y in thing:
print(x, y, x==y)
works as well.
For completeness, here's one possible implementation of the reversed-iterator:
class C_reversed_iterator:
def __init__(self, parent):
self.parent = parent
self.idx = len(parent.stuff) + 1
def __iter__(self):
return self
def __next__(self):
self.idx -= 1
if self.idx <= 0:
raise StopIteration
else:
return self.parent.stuff[self.idx - 1]
thing = C()
for x in reversed(thing):
for y in reversed(thing):
print(x, y, x==y)
Instead of defining your own iterators you could use generators. One way was already shown in the other answer:
class C:
def __init__(self):
self.stuff = ["A","B","C","D"]
def __iter__(self):
yield from self.stuff
def __reversed__(self):
yield from self.stuff[::-1]
or explicitly delegate to a generator function (that's actually equivalent to the above but maybe more clear that it's a new object that is produced):
def C_iterator(obj):
for item in obj.stuff:
yield item
def C_reverse_iterator(obj):
for item in obj.stuff[::-1]:
yield item
class C:
def __init__(self):
self.stuff = ["A","B","C","D"]
def __iter__(self):
return C_iterator(self)
def __reversed__(self):
return C_reverse_iterator(self)
Note: You don't have to implement the __reversed__ iterator. That was just meant as additional "feature" of the answer.
Your __iter__ is completely broken. Instead of actually making a fresh iterator on every call, it just resets some state on self and returns self. That means you can't actually have more than one iterator at a time over your object, and any call to __iter__ while another loop over the object is active will interfere with the existing loop.
You need to actually make a new object. The simplest way to do that is to use yield syntax to write a generator function. The generator function will automatically return a new iterator object every time:
class C(object):
def __init__(self):
self.stuff = ['A', 'B', 'C', 'D']
def __iter__(self):
for thing in self.stuff:
yield thing
I would like to have a function that can, optionally, return or yield the result.
Here is an example.
def f(option=True):
...
for...:
if option:
yield result
else:
results.append(result)
if not option:
return results
Of course, this doesn't work, I have tried with python3 and I always get a generator no matter what option value I set.
As far I have understood, python checks the body of the function and if a yield is present, then the result will be a generator.
Is there any way to get around this and make a function that can return or yield at will?
You can't. Any use of yield makes the function a generator.
You could wrap your function with one that uses list() to store all values the generator produces in a list object and returns that:
def f_wrapper(option=True):
gen = f()
if option:
return gen # return the generator unchanged
return list(gen) # return all values of the generator as a list
However, generally speaking, this is bad design. Don't have your functions alter behaviour like this; stick to one return type (a generator or an object) and don't have it switch between the two.
Consider splitting this into two functions instead:
def f():
yield result
def f_as_list():
return list(f())
and use either f() if you need the generator, and f_as_list() if you want to have a list instead.
Since list(), (and next() to access just one value of a generator) are built-in functions, you rarely need to use a wrapper. Just call those functions directly:
# access elements one by one
gen = f()
one_value = next(gen)
# convert the generator to a list
all_values = list(f())
What about this?
def make_f_or_generator(option):
def f():
return "I am a function."
def g():
yield "I am a generator."
if option:
return f
else:
return g
This gives you at least the choice to create a function or a generator.
class based approach
class FunctionAndGenerator:
def __init__(self):
self.counter = 0
def __iter__(self):
return self
# You need a variable to indicate if dunder next should return the string or raise StopIteration.
# Raising StopIteration will stop the loop from iterating more.
# You'll have to teach next to raise StopIteration at some point
def __next__(self):
self.counter += 1
if self.counter > 1 :
raise StopIteration
return f"I'm a generator and I've generated {self.counter} times"
def __call__(self):
return "I'm a function"
x = FunctionAndGenerator()
print(x())
for i in x:
print(i)
I'm a function
I'm a generator and I've generated 1 times
[Program finished]
I have a list of class objects and I'm using itertools.cycle() to continually loop through it.
class Letter:
def __init__(self,name):
self.name = name
self.att = 0
self.att1 = 0
Now I create the list of class objects
letterList = [Letter('a'),Letter('b'),Letter('c'),Letter('d')]
Then I make a sequence with itertools.cycle() like so:
seqList = itertools.cycle(letterList)
Anytime I want to go to the next item in the sequence and assign a value to one of the items attributes:
next(seqList).att = 1
But what if I want to assign another value to a different attribute? Is there another iterator method that will allow me to call the current item in the sequence like:
thisItem(seqList).att1 = 7
You can assign next() to a variable and access that variable.
this_one = next(seqList)
this_one.attr = value
If you just plan to infinitely loop over the same items, you can use a for loop (a break step is useful for getting out eventually though!):
for item in itertools.cycle(seqList):
item.attr1 = value1
item.attr2 = value2
if condition_met:
break
In most practical situations you simply use a variable, but just for fun, let's create a wrapper that would "remember" the last item (or N last items) for the given iterable:
from collections import deque
class recorded:
def __init__(self, it, size=1):
self.it = it
self.buf = deque(maxlen=size)
def __iter__(self):
return self
def next(self):
self.buf.append(next(self.it))
return self.buf[-1]
def last(self, n=0):
return self.buf[-1-n]
Now you can write:
it = recorded(itertools.cycle(letterList))
it.next()
it.last().something = 11
it.last().somethingElse = 111
or even:
it = recorded(itertools.cycle(letterList), 2)
it.next()
it.last().something = 11
it.next()
it.last(1).something = penultimate
The solution I found was:
next(seqList).att = 1
next(seqList)
next(seqList)
next(seqList)
next(seqList).att1 = 7
It's not pretty or elegant and it won't work very well for a larger set but it works for this situation and is very simple.