[Tutor] If you don't close file when writing, do bytes stay in memory?

Dave Angel davea at ieee.org
Sat Oct 10 13:02:08 CEST 2009

xbmuncher wrote:
> Which piece of code will conserve more memory?
>  I think that code #2 will because I close the file more often, thus freeing
> more memory by closing it.
> Am I right in this thinking... or does it not save me any more bytes in
> memory by closing the file often?
> Sure I realize that in my example it doesn't save much if it does... but I'm
> dealing with writing large files.. so every byte freed in memory counts.
> Thanks.
> CODE #1:
> def getData(): return '12345' #5 bytes
> f = open('file.ext', 'wb')
> for i in range(2000):
>     f.write(getData())
> f.close()
> CODE #2:
> def getData(): return '12345' #5 bytes
> f = open('file.ext', 'wb')
> for i in range(2000):
>     f.write(getData())
>     if i == 5:
>         f.close()
>         f = open('file.ext', 'ab')
>         i = 1
>     i = i + 1
> f.close()
You don't save a noticeable amount of memory usage by closing and 
immediately reopening the file.  The amount that the system buffers 
probably wouldn't depend on file size, in any case.  When dealing with 
large files, the thing to watch is how much of the data you've got in 
your own lists and dictionaries, not how much the file subsystem and OS 
are using.

But you have other issues in your code.

1) you don't say what version of Python you're using.  So I'll assume 
it's version 2.x.  If so, then range is unnecessarily using a lot of 
memory.  It builds a list of ints, when an iterator would do just as 
well.  Use xrange().  ( In Python 3.x, xrange() was renamed to be called 
range(). )  This may not matter for small values, but as the number gets 
bigger, so would the amount of wastage.

2) By using the same local for the for loop as for your "should I close" 
counter, you're defeating the logic.  As it stands, it'll only do the 
close() once.  Either rename one of these, or do the simpler test, of
      if i%5 == 0:
           f = open....

3) Close and re-open has three other effects.  One, it's slow.  Two, 
append-mode isn't guaranteed by the C standard to always position at the 
end (!).  And three, it flushes the data.  That can be a very useful 
result, in case the computer crashes while spending a long time updating 
a file.

I'd suggest sometimes doing a flush() call on the file, if you know 
you'll be spending a long time updating it.  But I wouldn't bother 
closing it.


More information about the Tutor mailing list