random writing access to a file in Python

Paul Rubin http
Sun Aug 27 13:57:47 EDT 2006


Claudio Grondi <claudio.grondi at freenet.de> writes:
> The Windows XP SP 2 '/> sort' (sorting of four Gigs of 20 byte records
> took 12 CPU and 18 usual hours) has, from what I could observe on the
> task manager, done the job in only two runs of 'copying' :

That is terrible; on a reasonably fast machine these days it should
take just a few minutes.

> > Is this some kind of production deployment problem you're working on,
> > or do you just have a big pile of data you need to sort once?
> The latter.
> One of the intermediate reasons behind doing it, is an attempt to get
> more and better intuitive understanding what are the hardware and
> software limits of brute force based approaches to solving problems in
> the area of AI, language processing and data compression.

This makes no sense; the methods you want are from the ancient days of
"data processing", nothing to do with AI.  Remember that the mainframe
computers of the 1960's that this stuff was developed on, had less
computing power and memory than a modern mobile phone.  But phone
companies, tax agencies, and so forth, still managed to run these big
sorting tasks on those mainframes in order to send people their phone
bills, identify the tax cheaters, etc.  There is rich and wonderful
history to it.



More information about the Python-list mailing list