# a=Time.now;(Time.parse(a.to\_s)-a)!=0 **URL:** \ **Category:** Ruby\ **Created:** [October 26, 2007, 6:04am UTC](https://www.ruby-forum.com/t/a-time-now-time-parse-a-to-s-a-0/118366 "2007-10-26T06:04:23Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![amin1426](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/4da419/32.png) [@amin1426](https://www.ruby-forum.com/u/amin1426)\ **Post date:** [October 26, 2007, 6:04am UTC](https://www.ruby-forum.com/t/a-time-now-time-parse-a-to-s-a-0/118366/1 "2007-10-26T06:04:23Z")
Hmm. Ooh yuck. Try this… $ ruby -r time -e ‘a=Time.now;p a-Time.parse(a.to\_s)’ 0.00848 $ ruby -r time -e ‘a=Time.now;p a-Time.parse(a.to\_s)’ 0.446575 $ ruby -r time -e ‘a=Time.now;p a-Time.parse(a.to\_s)’ 0.142796 ie. Time#to\_s doesn’t represent the full precision of the internal time format. Bit of a bummer if you want to round trip an exact timestamp onto disk and back. This does it right… $ ruby -e ‘a=Time.now;p a-Marshal::load(Marshal::dump(a))’ 0.0 …but look at whats on disk… ruby -e ‘a=Time.now;p Marshal::dump(a)’ “\004\bu:\tTime\rC\347\032\200z\343S\302” Eeew! Not exactly human friendly. The following is probably the most elegant way of exactly round tripping a time to disk and back in a human readable form? ruby -w -rtime -e ‘a=Time.now;b = a.xmlschema(6);p b;p a-Time.xmlschema(b)’ “2007-10-26T17:01:08.129059+13:00” 0.0 Ah well. John C. Phone : (64)(3) 358 6639 Tait Electronics Fax : (64)(3) 359 4632 PO Box 1645 Christchurch Email : removed\_email\_address@domain.invalid New Zealand ---
**Author:** ![amin1426](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/4da419/32.png) [@amin1426](https://www.ruby-forum.com/u/amin1426)\ **Post date:** [October 26, 2007, 6:17am UTC](https://www.ruby-forum.com/t/a-time-now-time-parse-a-to-s-a-0/118366/2 "2007-10-26T06:17:15Z")
On 26 Oct 2007, at 13:03, John C. wrote: > Time.xmlschema(b)’ > “2007-10-26T17:01:08.129059+13:00” > 0.0 You can use YAML if you only want Time. irb(main):002:0\> Time.now.to\_yaml =\> “— 2007-10-26 13:10:45.012414 +09:00\n” irb(main):003:0\> YAML.load(Time.now.to\_yaml) =\> Fri Oct 26 13:10:55 +0900 2007 irb(main):004:0\> t = Time.now =\> Fri Oct 26 13:13:05 +0900 2007 irb(main):005:0\> t - YAML.load(t.to\_yaml) =\> 0.0 Though there’s a bug with DateTime: irb(main):006:0\> DateTime.now.to\_yaml =\> “— !timestamp 2007-10-26T13:11:15+0900\n” irb(main):007:0\> YAML.load(DateTime.now.to\_yaml) =\> Wed Sep 19 10:11:26 +0900 2007 Alex G. Bioinformatics Center Kyoto University ---
**Author:** ![amin1426](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/4da419/32.png) [@amin1426](https://www.ruby-forum.com/u/amin1426)\ **Post date:** [October 26, 2007, 5:29pm UTC](https://www.ruby-forum.com/t/a-time-now-time-parse-a-to-s-a-0/118366/3 "2007-10-26T17:29:25Z")
John C. a écrit : > a-Time.xmlschema(b)’ Hi, It seems that Time objects are precise to the microsecond, but to\_s (and to\_i) limit the precision to one second. a= Time.now b = Time.parse(a.to\_s) a-b #=\> 0.875 a.usec #=\> 875000 I guess you could get a full representation of a Time object by combining to\_s (or to\_i) with #usec ---
**Author:** ![amin1426](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/4da419/32.png) [@amin1426](https://www.ruby-forum.com/u/amin1426)\ **Post date:** [October 26, 2007, 6:03pm UTC](https://www.ruby-forum.com/t/a-time-now-time-parse-a-to-s-a-0/118366/4 "2007-10-26T18:03:55Z")
On Oct 25, 2007, at 10:03 PM, John C. wrote: > ruby -w -rtime -e ‘a=Time.now;b = a.xmlschema(6);p b;p a- > Time.xmlschema(b)’ > “2007-10-26T17:01:08.129059+13:00” > 0.0 though i prefer #iso8601 to #xmlschema because even typing the word ‘xml’ makes my skin crawl 😉 a @ [http://codeforpeople.com/](http://codeforpeople.com/)