Significance of "start" parameter to string method "endswith"

Discussion in 'Python' started by =?utf-8?B?Qm9yaXMgRHXFoWVr?=, Apr 19, 2007.

1. =?utf-8?B?Qm9yaXMgRHXFoWVr?=Guest

Hello,

what is the use-case of parameter "start" in string's "endswith"
method? Consider the following minimal example:

a = "testing"
suffix="ing"
a.endswith(suffix, 2)

Significance of "end" is obvious. But not so for "start".

Let's assume the "end" parameter is not used - then the function
should simple check that the last "len(suffix)" characters of "a" are
equal to "ing", no matter where we start (the function does not *scan*
the string from the "start", does it?)
Only case where it would make difference is if we had start +
len(suffix) < len(a) (excuse possible "of-by-one" error
Then the function would never return True. But is there a real use
case when we would test for endswith like this? (knowing that it must
return false?)

Thanks for any ideas/experience.
Boris

=?utf-8?B?Qm9yaXMgRHXFoWVr?=, Apr 19, 2007

2. Larry BatesGuest

Boris DuÅ¡ek wrote:
> Hello,
>
> what is the use-case of parameter "start" in string's "endswith"
> method? Consider the following minimal example:
>
> a = "testing"
> suffix="ing"
> a.endswith(suffix, 2)
>
> Significance of "end" is obvious. But not so for "start".
>
> Let's assume the "end" parameter is not used - then the function
> should simple check that the last "len(suffix)" characters of "a" are
> equal to "ing", no matter where we start (the function does not *scan*
> the string from the "start", does it?)
> Only case where it would make difference is if we had start +
> len(suffix) < len(a) (excuse possible "of-by-one" error
> Then the function would never return True. But is there a real use
> case when we would test for endswith like this? (knowing that it must
> return false?)
>
> Thanks for any ideas/experience.
> Boris
>

Seems like a convenience I've never used.

>>> a="abcdef"
>>> a.endswith('cd',2,4)

True
>>> a[2:4].endswith('cd')

True

-Larry

Larry Bates, Apr 19, 2007

3. subscriber123Guest

On Apr 19, 3:58 pm, Boris DuÅ¡ek <> wrote:
> Hello,
>
> what is the use-case of parameter "start" in string's "endswith"
> method? Consider the following minimal example:
>
> a = "testing"
> suffix="ing"
> a.endswith(suffix, 2)
>
> Significance of "end" is obvious. But not so for "start".
>
> Let's assume the "end" parameter is not used - then the function
> should simple check that the last "len(suffix)" characters of "a" are
> equal to "ing", no matter where we start (the function does not *scan*
> the string from the "start", does it?)
> Only case where it would make difference is if we had start +
> len(suffix) < len(a) (excuse possible "of-by-one" error
> Then the function would never return True. But is there a real use
> case when we would test for endswith like this? (knowing that it must
> return false?)
>
> Thanks for any ideas/experience.
> Boris

Basically, this must be so in order for this to be Pythonic. This is
because it is an object oriented language, and functions can be passed
as arguments. Say, for example, you have the following function:

def foo(function,instance,param):
if function(instance,param,2,4):
return True
else: return False

The function must work whether you pass it
foo(str.endswith,"blaahh","ahh"), or
foo(str.startswith,"blaahh","aah"). This is a really bad example, but
it gets the point across that similar functions must have similar
parameters in order to be Pythonic.

I personally have never used the second or third parameters in this
function nor in str.startswith.

subscriber123, Apr 19, 2007
4. John MachinGuest

On Apr 20, 6:08 am, Larry Bates <> wrote:
> Boris DuÅ¡ek wrote:
> > Hello,

>
> > what is the use-case of parameter "start" in string's "endswith"
> > method? Consider the following minimal example:

>
> > a = "testing"
> > suffix="ing"
> > a.endswith(suffix, 2)

>
> > Significance of "end" is obvious. But not so for "start".

>
> > Let's assume the "end" parameter is not used - then the function
> > should simple check that the last "len(suffix)" characters of "a" are
> > equal to "ing", no matter where we start (the function does not *scan*
> > the string from the "start", does it?)
> > Only case where it would make difference is if we had start +
> > len(suffix) < len(a) (excuse possible "of-by-one" error
> > Then the function would never return True. But is there a real use
> > case when we would test for endswith like this? (knowing that it must
> > return false?)

Any or all of a, start and suffix can be variable. I can't see how we
could know that it must return false (except of course in the trivial
and useless case that they are all constant). IMHO it's much better in
general to let a string method do checks for corner cases (very
efficiently) than to waste brain cells trying to work out how to write
correct (but relatively very slow) Python code to avoid calling the
method.

>
> Seems like a convenience I've never used.
>
> >>> a="abcdef"
> >>> a.endswith('cd',2,4)

> True
> >>> a[2:4].endswith('cd')

>
> True
>

It's not just a convenience, and not just with endswith. In general:
astring[start:end].amethod(anarg) manifests the slice as a separate
temporary object, whereas astring.amethod(anarg, start, stop) does
not, and should be more efficient when one is looping over a long
string

Granted endswith's start arg is not wildly useful, but it's orthogonal
-- IMHO all string-processing functions should have a pair of start/
end args as a matter of course, unless neither start nor end is
useful.

Cheers,
John

John Machin, Apr 19, 2007
5. John MachinGuest

On Apr 20, 6:36 am, subscriber123 <> wrote:
> On Apr 19, 3:58 pm, Boris DuÅ¡ek <> wrote:
>
>
>
>
>
>
>
> > Hello,

>
> > what is the use-case of parameter "start" in string's "endswith"
> > method? Consider the following minimal example:

>
> > a = "testing"
> > suffix="ing"
> > a.endswith(suffix, 2)

>
> > Significance of "end" is obvious. But not so for "start".

>
> > Let's assume the "end" parameter is not used - then the function
> > should simple check that the last "len(suffix)" characters of "a" are
> > equal to "ing", no matter where we start (the function does not *scan*
> > the string from the "start", does it?)
> > Only case where it would make difference is if we had start +
> > len(suffix) < len(a) (excuse possible "of-by-one" error
> > Then the function would never return True. But is there a real use
> > case when we would test for endswith like this? (knowing that it must
> > return false?)

>
> > Thanks for any ideas/experience.
> > Boris

>
> Basically, this must be so in order for this to be Pythonic. This is
> because it is an object oriented language, and functions can be passed
> as arguments. Say, for example, you have the following function:
>
> def foo(function,instance,param):
> if function(instance,param,2,4):
> return True
> else: return False

Perhaps
return function(instance, param, 2, 4)
would have a higher pythonicity index

>
> The function must work whether you pass it
> foo(str.endswith,"blaahh","ahh"), or
> foo(str.startswith,"blaahh","aah"). This is a really bad example, but
> it gets the point across that similar functions must have similar
> parameters in order to be Pythonic.
>
> I personally have never used the second or third parameters in this
> function nor in str.startswith.

John Machin, Apr 19, 2007