file.close()

[Bryan]

i'm curious to know how others handle the closing of files. i seem to
always end up with this pattern even though i rarely see others do it.

f1 = file('file1')
try:
# process f1
finally:
f1.close()

which explicitly closes f1

or for multiple files:

f1 = file('file1')
try:
f2 = file('file2')
try:
# process f1 & f2
finally:
f2.close()
finally:
f1.close()

which explicitly closes f1 & f2
any exceptions opening f1 or f2 is handled outside of this structure or is
allowed to fall out of the program. i'm aware that files will automatically
be closed when the process exits. i'm just curious how others do this for
small scripts and larger programs. is what i'm doing is overkill?

thanks,

bryan

[Erik Max Francis]

Re: file.close()

Bryan wrote:
[color=blue]
> which explicitly closes f1 & f2
> any exceptions opening f1 or f2 is handled outside of this structure
> or is
> allowed to fall out of the program. i'm aware that files will
> automatically
> be closed when the process exits. i'm just curious how others do this
> for
> small scripts and larger programs. is what i'm doing is overkill?[/color]

No, not at all; it's safe and portable. Python the language does not
specify when objects get reclaimed, although CPython the implementation
does it promptly. Use of external resources -- which should be released
as soon as you're done with them -- are best done in try/finally
clauses.

--
Erik Max Francis && max@alcyone.com && http://www.alcyone.com/max/
__ San Jose, CA, USA && 37 20 N 121 53 W && &tSftDotIotE
/ \ It is fatal to enter any war without the will to win it.
\__/ Douglas MacArthur

[Ben Finney]

Re: file.close()

On Wed, 23 Jul 2003 20:29:27 -0700, Erik Max Francis wrote:[color=blue]
> Bryan wrote:[color=green]
>> is what i'm doing is overkill?[/color]
>
> Use of external resources -- which should be released as soon as
> you're done with them -- are best done in try/finally clauses.[/color]

It seems that nesting the 'try' clauses doesn't scale well. What if
fifty files are opened? Must the nesting level of the 'try' clauses be
fifty also, to close them promptly?

--
\ "God forbid that any book should be banned. The practice is as |
`\ indefensible as infanticide." -- Dame Rebecca West |
_o__) |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B

[Erik Max Francis]

Re: file.close()

Ben Finney wrote:
[color=blue]
> It seems that nesting the 'try' clauses doesn't scale well. What if
> fifty files are opened? Must the nesting level of the 'try' clauses
> be
> fifty also, to close them promptly?[/color]

If you're manipulating fifty files in one block, presumably you're doing
so in a uniform way:

allFiles = [...]
try:
...
finally:
for eachFile in allFiles:
eachFile.close( )

--
Erik Max Francis && max@alcyone.com && http://www.alcyone.com/max/
__ San Jose, CA, USA && 37 20 N 121 53 W && &tSftDotIotE
/ \ Man is a hating rather than a loving animal.
\__/ Rebecca West

[Carl Banks]

Re: file.close()

Ben Finney wrote:[color=blue]
> On Wed, 23 Jul 2003 20:29:27 -0700, Erik Max Francis wrote:[color=green]
>> Bryan wrote:[color=darkred]
>>> is what i'm doing is overkill?[/color]
>>
>> Use of external resources -- which should be released as soon as
>> you're done with them -- are best done in try/finally clauses.[/color]
>
> It seems that nesting the 'try' clauses doesn't scale well. What if
> fifty files are opened? Must the nesting level of the 'try' clauses be
> fifty also, to close them promptly?[/color]

If you need 50 open files, you almost certainly want to have a way of
organizing them. Probably they'll be in a list or dictionary. So, if
they're in a list, for example, you can do this:


filelist = []
try:
filelist.append (open(filename[0]))
filelist.append (open(filename[1]))
...
do_something(fi lelist)
finally:
for f in filelist:
f.close()


--
CARL BANKS

[Ben Finney]

Re: file.close()

Bryan (original poster) wrote:[color=blue]
> f1 = file('file1')
> try:
> f2 = file('file2')
> try:
> # process f1 & f2
> finally:
> f2.close()
> finally:
> f1.close()[/color]


On Wed, 23 Jul 2003 21:12:34 -0700, Erik Max Francis wrote:[color=blue]
> Ben Finney wrote:[color=green]
>> It seems that nesting the 'try' clauses doesn't scale well. What if
>> fifty files are opened? Must the nesting level of the 'try' clauses
>> be fifty also, to close them promptly?[/color]
>
> If you're manipulating fifty files in one block, presumably you're
> doing so in a uniform way:
>
> allFiles = [...]
> try:
> ...
> finally:
> for eachFile in allFiles:
> eachFile.close( )[/color]

This doesn't match Bryan's nested structure above, which you blessed as
not "overkill" (in his words). It was this that I considered a
poorly-scaling structure, or "overkill" since only one 'try' block is
required. Do you disagree?

--
\ "Too many Indians spoil the golden egg." -- Sir Joh |
`\ Bjelke-Petersen |
_o__) |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B

[Erik Max Francis]

Re: file.close()

Ben Finney wrote:
[color=blue]
> This doesn't match Bryan's nested structure above, which you blessed
> as
> not "overkill" (in his words). It was this that I considered a
> poorly-scaling structure, or "overkill" since only one 'try' block is
> required. Do you disagree?[/color]

It uses try/finally to secure the closing of many files in a timely
manner. In that sense, it certainly fits the pattern. It doesn't have
the same nested pattern, but try/finally isn't at issue here. If you
had code which opened 50 files and looked like:

fileOne = file(...)
fileTwo = file(...)
fileThree = file(...)
...
fileFortyNine = file(...)
fileFifty = file(...)

I would say you are doing something wrong. A more systematic handling
of many, many files is indicated whether or not you're using the
try/finally idiom to ensure files get closed in a timely manner.

--
Erik Max Francis && max@alcyone.com && http://www.alcyone.com/max/
__ San Jose, CA, USA && 37 20 N 121 53 W && &tSftDotIotE
/ \ Together we can take this one day at a time
\__/ Sweetbox

[Bryan]

Re: file.close()

>[color=blue]
> If you need 50 open files, you almost certainly want to have a way of
> organizing them. Probably they'll be in a list or dictionary. So, if
> they're in a list, for example, you can do this:
>
>
> filelist = []
> try:
> filelist.append (open(filename[0]))
> filelist.append (open(filename[1]))
> ...
> do_something(fi lelist)
> finally:
> for f in filelist:
> f.close()
>[/color]


erik, carl... thanks... this is exactly what i was looking for

bryan

[Ben Finney]

Re: file.close()

On Thu, 24 Jul 2003 05:20:15 GMT, Bryan wrote:[color=blue][color=green]
>> filelist = []
>> try:
>> filelist.append (open(filename[0]))
>> filelist.append (open(filename[1]))
>> ...
>> do_something(fi lelist)
>> finally:
>> for f in filelist:
>> f.close()[/color]
>
> erik, carl... thanks... this is exactly what i was looking for[/color]

The only substantial difference I see between this and what you
originally posted, is that there is only one 'try...finally' block for
all the file open/close operations. Is this what you were wanting
clarified?

--
\ "I spent all my money on a FAX machine. Now I can only FAX |
`\ collect." -- Steven Wright |
_o__) |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B

[Bryan]

Re: file.close()


"Ben Finney" <bignose-hates-spam@and-benfinney-does-too.id.au> wrote in
message news:slrnbhusl6 .13c.bignose-hates-spam@iris.polar .local...[color=blue]
> On Thu, 24 Jul 2003 05:20:15 GMT, Bryan wrote:[color=green][color=darkred]
> >> filelist = []
> >> try:
> >> filelist.append (open(filename[0]))
> >> filelist.append (open(filename[1]))
> >> ...
> >> do_something(fi lelist)
> >> finally:
> >> for f in filelist:
> >> f.close()[/color]
> >
> > erik, carl... thanks... this is exactly what i was looking for[/color]
>
> The only substantial difference I see between this and what you
> originally posted, is that there is only one 'try...finally' block for
> all the file open/close operations. Is this what you were wanting
> clarified?
>[/color]

well, it wasn't exactly clarification as much as seeing another more
scalable solution. i know a lot of people don't explicitly close files in
python, but i always do even for small scripts.

thanks again,

bryan

[Ben Finney]

Re: file.close()

On Wed, 23 Jul 2003 22:19:07 -0700, Erik Max Francis wrote:[color=blue]
> Ben Finney wrote:[color=green]
>> This doesn't match Bryan's nested structure above, which you blessed
>> as not "overkill" (in his words).[/color]
> It doesn't have the same nested pattern, but try/finally isn't at
> issue here.[/color]

Judging by Bryan's responses elsewhere in this thread, the multiple
nested 'try...finally' is indeed what he was asking about. The question
seems to be answered now.

--
\ "Those who will not reason, are bigots, those who cannot, are |
`\ fools, and those who dare not, are slaves." -- "Lord" George |
_o__) Gordon Noel Byron |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B

[Francois Pinard]

Re: file.close()

[Bryan]
[color=blue]
> I'm curious to know how others handle the closing of files. [...] I'm
> aware that files will automatically be closed when the process exits.[/color]

For one, I systematically avoid cluttering my code with unneeded `close'.
The advantages are simplicity and legibility, both utterly important to me.

However, I do understand that if I ever have to move a Python script
to Jython, I will have to revise my scripts for adding the clutter I am
sparing today. I'm quite accepting to do that revision if this occurs.
Until then, I prefer keeping my scripts as neat as possible.

For me, explicitely closing a file, for which the only reference is about
to disappear through function exiting, would be very similar to using
`del' on any variable I happened to use in that function: gross overkill...

The only reason to call `close' explicitly is when there is a need to close
prematurely. Absolutely no doubt that such needs exist at times. But
closing all the time "just in case" is symptomatic of unsure programming.
Or else, it is using Python while still thinking in other languages.

--
François Pinard http://www.iro.umontreal.ca/~pinard

[Bryan]

Re: file.close()


"Francois Pinard" <pinard@iro.umo ntreal.ca> wrote in message
news:mailman.10 59052897.16526. python-list@python.org ...[color=blue]
> [Bryan]
>[color=green]
> > I'm curious to know how others handle the closing of files. [...] I'm
> > aware that files will automatically be closed when the process exits.[/color]
>
> For one, I systematically avoid cluttering my code with unneeded `close'.
> The advantages are simplicity and legibility, both utterly important to[/color]
me.[color=blue]
>
> However, I do understand that if I ever have to move a Python script
> to Jython, I will have to revise my scripts for adding the clutter I am
> sparing today. I'm quite accepting to do that revision if this occurs.
> Until then, I prefer keeping my scripts as neat as possible.
>
> For me, explicitely closing a file, for which the only reference is about
> to disappear through function exiting, would be very similar to using
> `del' on any variable I happened to use in that function: gross[/color]
overkill...[color=blue]
>
> The only reason to call `close' explicitly is when there is a need to[/color]
close[color=blue]
> prematurely. Absolutely no doubt that such needs exist at times. But
> closing all the time "just in case" is symptomatic of unsure programming.
> Or else, it is using Python while still thinking in other languages.
>
> --
> François Pinard http://www.iro.umontreal.ca/~pinard
>[/color]

you are correct this is so awesome... at least for me it is... now i remove
a lot of my clutter too :) i just did some tests on windows by trying to
delete file1 at the command prompt when raw_input is called.

f = file('file1')
raw_input('paus e')

### the file is NOT closed


file('file1')
raw_input('paus e')

### the file IS closed


f = file('file1')
del f
raw_input('paus e')

### the file IS closed


def foo():
f = file('file1')

foo()
raw_input('paus e')

### the file IS closed



can you explain to me how the file gets closed? i'm sure that garbage
collection hasn't happed at the point that i call raw_input. it must have
something to do with the reference count of the file object. does python
immediately call close for you when the reference count goes to zero? i
want the dirty details...

thanks,

bryan

[Ulrich Petri]

Re: file.close()


"Francois Pinard" <pinard@iro.umo ntreal.ca> schrieb im Newsbeitrag
news:mailman.10 59052897.16526. python-list@python.org ...[color=blue]
> [Bryan]
>[color=green]
> > I'm curious to know how others handle the closing of files. [...] I'm
> > aware that files will automatically be closed when the process exits.[/color]
>
> For one, I systematically avoid cluttering my code with unneeded `close'.
> The advantages are simplicity and legibility, both utterly important to[/color]
me.[color=blue]
>
> However, I do understand that if I ever have to move a Python script
> to Jython, I will have to revise my scripts for adding the clutter I am
> sparing today. I'm quite accepting to do that revision if this occurs.
> Until then, I prefer keeping my scripts as neat as possible.
>
> For me, explicitely closing a file, for which the only reference is about
> to disappear through function exiting, would be very similar to using
> `del' on any variable I happened to use in that function: gross[/color]
overkill...[color=blue]
>
> The only reason to call `close' explicitly is when there is a need to[/color]
close[color=blue]
> prematurely. Absolutely no doubt that such needs exist at times. But
> closing all the time "just in case" is symptomatic of unsure programming.
> Or else, it is using Python while still thinking in other languages.
>[/color]

fire up your python and type: "import this" read the output and rethink your
codint techniques

Ciao Ulrich

[Erik Max Francis]

Re: file.close()

Bengt Richter wrote:
[color=blue]
> 1) Is f.done() really necessary? I.e., doesn't an explicit del f take
> care of it
> if the object has been coded with a __del__ method? I.e., the idea
> was to get
> the effect of CPython's immediate effective del on ref count going
> to zero, right?[/color]

It wouldn't if there were circular references at that point. If you're
going to have some kind of `with' structure that constraints lifetimes,
I'd think you'd probably want something more concrete than just object
deletion; you'd want to make sure a "Stop whatever you were doing now"
method were present and called. But maybe that really depends on the
primary thing that the `with' construct would be used for.

--
Erik Max Francis && max@alcyone.com && http://www.alcyone.com/max/
__ San Jose, CA, USA && 37 20 N 121 53 W && &tSftDotIotE
/ \ Love is when you wake up in the morning and have a big smile.
\__/ Anggun