How to stop iteration with __iter__() ?

[ssecorp]

I want a parse a file of the format:
movieId
customerid, grade, date
customerid, grade, date
customerid, grade, date
etc.

so I could do with open file as reviews and then for line in reviews.

but first I want to take out the movie id so I use an iterator.

then i want to iterate through all the rows, but how can I do:
while movie_iter != None:

because that doesn't work, itraises an exception, StopItreation, which
according to the documentation it should. But catching an exception
can't be the standard way to stop iterating right?

[Jeff]

On Aug 19, 7:39*am, ssecorp <[e-mail]:

I want a parse a file of the format:
movieId
customerid, grade, date
customerid, grade, date
customerid, grade, date
etc.

so I could do with open file as reviews and then for line in reviews.

but first I want to take out the movie id so I use an iterator.

then i want to iterate through all the rows, but how can I do:
while movie_iter != None:

because that doesn't work, itraises an exception, StopItreation, which
according to the documentation it should. But catching an exception
can't be the standard way to stop iterating right?

No. If you are defining the iterator, then the iterator raises
StopIteration when it is complete. It is not an exception you
normally catch; you use a for loop to iterate, and the machinery of
the for loop uses the StopIteration exception to tell it when to
break. A file-like object is itself a line iterator, so:

def review_iter(file_obj):
for line in file_obj:
if matches_some_pattern(line):
yield line

Then, to use it:

file_obj = open('/path/to/some/file.txt')
try:
for review in review_iter(file_obj):
do_something(review)
finally:
file_obj.close()

[Jon Clements]

On Aug 19, 12:39 pm, ssecorp <[e-mail]:

I want a parse a file of the format:
movieId
customerid, grade, date
customerid, grade, date
customerid, grade, date
etc.

so I could do with open file as reviews and then for line in reviews.

but first I want to take out the movie id so I use an iterator.

then i want to iterate through all the rows, but how can I do:
while movie_iter != None:

because that doesn't work, itraises an exception, StopItreation, which
according to the documentation it should. But catching an exception
can't be the standard way to stop iterating right?

Given your example, why not just open the file and call .next() on it
to discard the first row, then iterate using for as usual?

hth
Jon

[Leo Jay]

On Tue, Aug 19, 2008 at 7:39 PM, ssecorp <ci**********@gmail.comwrote:

I want a parse a file of the format:
movieId
customerid, grade, date
customerid, grade, date
customerid, grade, date
etc.

so I could do with open file as reviews and then for line in reviews.

but first I want to take out the movie id so I use an iterator.

then i want to iterate through all the rows, but how can I do:
while movie_iter != None:

because that doesn't work, itraises an exception, StopItreation, which
according to the documentation it should. But catching an exception
can't be the standard way to stop iterating right?

you can use for loop:
for line in movie_iter:
...

--
Best Regards,
Leo Jay

[Terry Reedy]

ssecorp wrote:

I want a parse a file of the format:
movieId
customerid, grade, date
customerid, grade, date
customerid, grade, date
etc.

so I could do with open file as reviews and then for line in reviews.

but first I want to take out the movie id so I use an iterator.

then i want to iterate through all the rows, but how can I do:
while movie_iter != None:

because that doesn't work, itraises an exception, StopItreation, which
according to the documentation it should. But catching an exception
can't be the standard way to stop iterating right?

Catching StopIteration *is* the usual way to stop iterating with an
iterator. Using while, one must be explicit:

it = iter(iterable)
try:
while True:
item = next(iter) # 2.6,3.0
f(item)
except StopIteration:
pass

but the main reason to write the above is to show the advantage of
simply writing the equivalent (and implicit)

for item in iterable:
f(item)

;-)

In your case, the standard Python idiom, as Jon said, is

it = iter(iterable)
next(it) # 2.6, 3.0
for for item in iterable:
f(item)

The alternative is a flag variable and test

first = True
for for item in iterable:
if first:
first = False
else:
f(item)

This takes two more lines and does an unnecessary test for every line
after the first. But this approach might be useful if, for instance,
you needed to skip every other line (put 'first = True' after f(item)).

Terry Jan Reedy

[John Machin]

On Aug 20, 5:06 am, Terry Reedy <[e-mail]:

In your case, the standard Python idiom, as Jon said, is

it = iter(iterable)
next(it) # 2.6, 3.0
for for item in iterable:
f(item)

or, perhaps, for completeness/paranoia/whatever:

it = iter(iterable)
try:
headings = it.next() # < 2.5
except StopIteration:
# code to handle empty <iterable>
for item etc etc

The alternative is a flag variable and test

first = True
for for item in iterable:
if first:
first = False
else:
f(item)

This takes two more lines and does an unnecessary test for every line
after the first. But this approach might be useful if, for instance,
you needed to skip every other line (put 'first = True' after f(item)).

and change its name from 'first' to something more meaningful ;-)

Cheers,
John

[MRAB]

On Aug 20, 12:11*am, John Machin <[e-mail]:

On Aug 20, 5:06 am, Terry Reedy <[e-mail]:

In your case, the standard Python idiom, as Jon said, is

it = iter(iterable)
next(it) # 2.6, 3.0
for for item in iterable:
* *f(item)

or, perhaps, for completeness/paranoia/whatever:

it = iter(iterable)
try:
* *headings = it.next() # < 2.5
except StopIteration:
* *# code to handle empty <iterable>
for item etc etc

I think it needs to be:

it = iter(iterable)
try:
headings = it.next() # < 2.5
except StopIteration:
# code to handle empty <iterable>
else:
for item etc etc

because you don't want to iterate over the remainder if it has already
stopped yielding! :-)

The alternative is a flag variable and test

first = True
for for item in iterable:
* *if first:
* * *first = False
* *else:
* * *f(item)

This takes two more lines and does an unnecessary test for every line
after the first. *But this approach might be useful if, for instance,
you needed to skip every other line (put 'first = True' after f(item)).

and change its name from 'first' to something more meaningful ;-)

Cheers,
John

[John Machin]

On Aug 20, 7:56*pm, MRAB <[e-mail]:

On Aug 20, 12:11*am, John Machin <[e-mail]:

or, perhaps, for completeness/paranoia/whatever:

it = iter(iterable)
try:
* *headings = it.next() # < 2.5
except StopIteration:
* *# code to handle empty <iterable>
for item etc etc

I think it needs to be:

it = iter(iterable)
try:
* * headings = it.next() # < 2.5
except StopIteration:
* * # code to handle empty <iterable>
else:
* * for item etc etc

because you don't want to iterate over the remainder if it has already
stopped yielding! :-)

If it has stopped yielding, the remainder is nothing/null/empty, isn't
it?
In any case "code to handle empty <iterable>" is likely to end with a
raise or return statement.

Cheers,
John

[MRAB]

On Aug 20, 11:27*am, John Machin <[e-mail]:

On Aug 20, 7:56*pm, MRAB <[e-mail]:

On Aug 20, 12:11*am, John Machin <[e-mail]:

or, perhaps, for completeness/paranoia/whatever:

it = iter(iterable)
try:
* *headings = it.next() # < 2.5
except StopIteration:
* *# code to handle empty <iterable>
for item etc etc

I think it needs to be:

it = iter(iterable)
try:
* * headings = it.next() # < 2.5
except StopIteration:
* * # code to handle empty <iterable>
else:
* * for item etc etc

because you don't want to iterate over the remainder if it has already
stopped yielding! :-)

If it has stopped yielding, the remainder is nothing/null/empty, isn't
it?
In any case "code to handle empty <iterable>" is likely to end with a
raise or return statement.

So it's defined behaviour that an exhausted iterable will always raise
StopIteration no matter how many times it's asked for the next value?

[Peter Otten]

MRAB wrote:

So it's defined behaviour that an exhausted iterable will always raise
StopIteration no matter how many times it's asked for the next value?

That is not enforced. However, quoting
http://www.python.org/dev/peps/pep-0234/

"""
Iterator implementations (in C or in Python) should guarantee that
once the iterator has signalled its exhaustion, subsequent calls
to tp_iternext or to the next() method will continue to do so.
"""

Peter

[Steven D'Aprano]

On Wed, 20 Aug 2008 06:16:17 -0700, MRAB wrote:

So it's defined behaviour that an exhausted iterable will always raise
StopIteration no matter how many times it's asked for the next value?

Be careful -- an iterable is not the same as an iterator.

Iterables are anything that you can iterate over. This includes strings,
lists, tuples, dicts and, yes, iterators.

But not all iterables are iterators. Only iterators raise StopIteration.
The defined behaviour of iterators is that once exhausted, they will
always raise StopIteration.

Basically, Python will do the right thing in a for loop regardless of
whether you pass an iterator or some other type of iterable. You don't
have to worry about the mechanics of it, but if you care, it works
something like this:

When Python executes a for-loop:

for item in obj:
BLOCK

it does something kind of like this:
if hasattr(obj, "next"):
# Looks like an iterator.
while True:
try:
item = obj.next()
except StopIteration:
break
execute BLOCK
elif hasattr(obj, "__getitem__"):
# Looks like a string, or list, or similar.
i = 0
while True:
try:
item = obj[i]
except IndexError:
break
execute BLOCK
i += 1
else:
raise TypeError('object not iterable')

--
Steven