Am Donnerstag, 8. Dezember 2016 14:47:31 UTC+1 schrieb DFS: > On 12/8/2016 12:16 AM, Steven D'Aprano wrote: > > On Thursday 08 December 2016 03:15, DFS wrote: > > > >> dts= ['10-Mar-1998', > >> '20-Aug-1997', > >> '06-Sep-2009', > >> '23-Jan-2010', > >> '12-Feb-2010', > >> '05-Nov-2010', > >> '03-Sep-2009', > >> '07-Nov-2014', > >> '08-Mar-2013'] > >> > >> Of course, the naive: > >> min(dates) = '03-Sep-2009' > >> max(dates) = '23-Jan-2010' > >> is wrong. > >> > >> Not wanting to use any date parsing libraries, I came up with: > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > > > > That's where you have gone wrong. By not using a date parsing library which > > works and has been tested, you have to write your own dodgy and possibly > > buggy > > parsing routine yourself. > > > > Why reinvent the wheel? > > > Because the "wheel" is a pain in the ass. >
Why? And why do you use this wording? > -------------------------------------------------------------- > import time > dts=['10-Mar-1908','20-Aug-1937','06-Sep-1969','23-Jan-1952'] > def date_to_seconds(string): > return time.mktime(time.strptime(string, '%d-%b-%Y')) > print min(dts, key=date_to_seconds) > -------------------------------------------------------------- > > OverflowError: mktime argument out of range > > (snip) > > I like that flexibility, but mktime is apparently useless for dates > prior to 'the epoch'. With a little more experience in Python programming you should have discovered that time.mktime is not even required to do your calculations. Please remove time.mktime from the function date_to_seconds and rename the function to date_to_timestruct. -- Paolo -- https://mail.python.org/mailman/listinfo/python-list
