|Perl: the Markov chain saw|
Re: Modify csv file.by marinersk (Curate)
|on Nov 06, 2012 at 07:16 UTC||Need Help??|
Excellent replies in this thread but I fear none communicate what appears to be the primary faulty assumption you seem to making.
Modifying a file in place is usually left to the realm of binary file I/O for a list of reasons which could fill a chapter in CompSci 110.
CSV files are notoriously left to the realm of sequential text file processing, for a list of reasons that could at least fill a Wiki page somewhere.
You are, essentially, trying to adjust a sequential text file using a binary file approach. This is fraught with difficulties (and as one clever poster demonstrates, fun ones to solve) and for the sake of maintainable code, should probably be avoided.
One approach which might suit you better (and it might not, it is an engineering decision you will have to make) is to use the slurp-digest-spew model. Read the whole CSV into an array. Adjust the array in memory. Write the whole array out to the original file. And pray there is not a power outage in the middle of the write loop or you risk corrupting your data file.
I sincerely hope this, added to the immensely useful information already provided, may shed some light on why the thing you are attempting is giving you such headaches.
You are trying to fit a square peg into a round hole, which upon occasion may require you to grind away at the corners of the peg. This may not constitute good software engineering for your project when all is said and done.
As always, however, that is your call.