Use numbered_filename('sample-*.xml')
Python does not have a routine to find the next filename in a numbered sequence, so I wrote a simple module (see below). Usage is:
from numbered_filename import numbered_filename
fn = numbered_filename('sample-*.xml')
fh = open(fn, 'w')
rs = [blockresult]
fh.writelines(rs)
fh.close()
The first time the code is run, the output will be to sample-000.xml. The next run will write to sample-001.xml, then sample-002.xml, and so on. Each subsequent run increments the sequence number by one.
The module code
Save the following code into a file called numbered_filename.py.
"""Provide a function for creating sequentially incremented filenames
based upon a simple template in which an asterisk is replaced with a
number. The filesystem is checked for existing files that match the
template and the returned filename's sequence number is always one
greater than the maximum found.
"""
import glob
if __debug__:
import os
def numbered_filename(template :str ='', width :int =3) -> str:
"""Return the next filename in an incrementing sequence by adding
one to the current largest number in existing filenames.
template :str: a string with an asterisk in it representing where
the numbers are placed. ('foo-*.txt').
width :int: optional minimum number of digits to zero-pad the
sequence to. Defaults to 3 ('000', '001', '002', ...)
Example usage:
from numbered_filename import numbered_filename
newfile = numbered_filename('foo-*.txt')
with open(newfile, 'w') as outfile:
outfile.write("Bob's your uncle!")
Given a filename template with an asterisk in it, such as
'foo-*.txt', returns the same filename with the asterisk replaced
with the next number in the sequence, such as 'foo-007.txt'. If no
prior file exists, numbering starts at zero ('foo-000.txt').
The number will be left-padded with zeroes to contain at least
three digits, unless the optional 'width' argument is given.
Zero-padding can be disabled with 'width=0'. For example,
'numbered_filename("hackerb*", width=0)' might return 'hackerb9'.
Note that 'width' is a minimum and more digits will be used if
necessary. (E.g., 'foo-1000.txt').
Regardless of the 'width' setting, existing filenames need not be
zero-padded to be recognized. For example, if a directory has the
file 'foo-6.txt', the next filename will be 'foo-007.txt'.
This routine always return the next higher number after any
existing file, even if a lower number is available. For example,
in a directory containing only 'foo-099.txt', the next file would
be 'foo-100.txt', despite 'foo-000' through '-098.txt' being possible.
Peculiar Circumstances: If the template is the empty string (''),
then the output will simply be a sequence number ('007'). If the
template contains no asterisks ('foo'), then the number is
appended to the end of the filename ('foo007'). If more than one
asterisk is used ('*NSYNC*.txt'), then only the rightmost asterisk
is replaced with a number ('*NSYNC007.txt'). All others asterisks
are kept as literal '*' in the filename.
CAVEAT: While the code attempts to return an unused filename, it
is not guaranteed as there is a fairly obvious race condition. To
avoid it, processes writing to the same directory concurrently
must not use the same template. Do not use this to create temp
files in a directory where an adversary may have write access,
such as /tmp -- instead use 'mkstemp'.
"""
if not isinstance(template, str):
raise TypeError("numbered_filename() requires a string as a template, such as foo-*.txt")
(filename, asterisk, extension) = template.rpartition('*')
if not asterisk:
(filename, extension) = (extension, filename)
template=f'{filename}*'
try:
files = [int(f.lstrip(filename).rstrip(extension))
for f in glob.glob(template)
if f.lstrip(filename).rstrip(extension).isdigit()]
num = sorted(files)[-1]
except (IndexError, ValueError):
num = -1
num = num + 1
spec = f'0>{width}'
numstr = format(num, spec)
if __debug__:
result = filename + numstr + extension
if os.path.exists(result):
raise AssertionError(f'Error: "{result}" already exists. Race condition or bug?')
return filename + numstr + extension
Race condition warning
This module solves the problem described in the question, however, it makes no pretense of being secure. If your program is creating temporary files in a directory which an adversary has write access to, such as /tmp, you should use mkstemp() instead of numbered_filename().
