On 10/08/2020 05:23, Dennis Lee Bieber wrote:
On Sun, 9 Aug 2020 11:50:51 +1200, dn via Python-list
<python-list@python.org> declaimed the following:
To be a logomach, let's talk about "update":-
May I advise that a 'good practice' would be to create a new file, and
thus be able to (also) maintain the old version as a 'backup'.
(also: advantage when debugging/testing logic!)
Per the documentation, this is exactly what the "inplace=True" does! It
renames the original file, then writes the output under the original name.
Keeping the backup after processing requires providing an explicit
"backup=.ext" option, otherwise it uses .bak and deletes the file when
processing completes.
Apologies for incomplete answer.
I have finally remembered why I (have long) steer-clear of fileinput:
a) criticism from Martijn Pieters
https://www.zopatista.com/python/2013/11/26/inplace-file-rewriting/
b) I can't recall 'my' last application requiring/desiring a flat-file
that wasn't JSON or YAML.
The above article is (2013) old. I notice that whilst the github
received a trivial update one month back, that the PyPI entry is almost
two years old.
I (am biased, admittedly) still maintain that for anything more than the
trivial, a direct-access mechanism is the tool-for-the-job, and because
I think they are 'easy', an RDBMS. YMMV!
--
Regards =dn
--
https://mail.python.org/mailman/listinfo/python-list
