Comments on: UNIX Time and UNIX Timestamps http://www.lifeaftercoffee.com/2006/04/18/unix-time-and-unix-timestamps/ because I don't believe in life before coffee... Wed, 23 Nov 2011 03:01:12 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Mike http://www.lifeaftercoffee.com/2006/04/18/unix-time-and-unix-timestamps/comment-page-1/#comment-379966 Mike Sat, 27 Aug 2011 22:40:38 +0000 http://www.lifeaftercoffee.com/2006/04/18/unix-time-and-unix-timestamps/#comment-379966 I wrote up a handy utility for developers that need to convert unix timestamps frequently: http://www.unixstamp.com It's a wrapper for the strtotime() function so you can convert pretty much anything: unixstamp.com/1234567 unixstamp.com/last+tuesday unixstamp.com/2012/10/15 Hope you find it useful! I wrote up a handy utility for developers that need to convert unix timestamps frequently: http://www.unixstamp.com

It’s a wrapper for the strtotime() function so you can convert pretty much anything:
unixstamp.com/1234567
unixstamp.com/last+tuesday
unixstamp.com/2012/10/15

Hope you find it useful!

]]> By: Jon http://www.lifeaftercoffee.com/2006/04/18/unix-time-and-unix-timestamps/comment-page-1/#comment-4806 Jon Mon, 04 Sep 2006 17:55:47 +0000 http://www.lifeaftercoffee.com/2006/04/18/unix-time-and-unix-timestamps/#comment-4806 Jim, Yes, if multiple processes ask for the UNIX timestamp durring the same second they will get the exact same timestamp. Because of this it is almost always a bad idea to use a date or timestamp for a primary key. If this is a database you will want to set up an auto-increment in one way or another. The methods vary by database, but if it's Oracle I have directions posted here: http://www.lifeaftercoffee.com/2006/02/17/how-to-create-auto-increment-columns-in-oracle/ Hope this helps. Jim,

Yes, if multiple processes ask for the UNIX timestamp durring the same second they will get the exact same timestamp. Because of this it is almost always a bad idea to use a date or timestamp for a primary key.

If this is a database you will want to set up an auto-increment in one way or another. The methods vary by database, but if it’s Oracle I have directions posted here:

http://www.lifeaftercoffee.com/2006/02/17/how-to-create-auto-increment-columns-in-oracle/

Hope this helps.

]]> By: Jim Cronin http://www.lifeaftercoffee.com/2006/04/18/unix-time-and-unix-timestamps/comment-page-1/#comment-4527 Jim Cronin Thu, 31 Aug 2006 15:16:17 +0000 http://www.lifeaftercoffee.com/2006/04/18/unix-time-and-unix-timestamps/#comment-4527 I have a question for anyone, regarding the use of Unix Epoc time as a "guaranteed, unique" record ID. I manage a system in a non-Unix environment that utilizes a similar "time stamp", using an internal 5-digit date, concatenated with seconds-since-midnight. If I wish to use this stamp as a record ID, I have to expect that other requests are being made for the time-date stamp, and may find an alpha character appended (1411035660a). I have actual examples of this scenario where the letter "p" has been appended, meaning 15 other applications were asking for time-date stamp, simultaneously. At the speed of today's processors, this can be expected. In another test, I wrote the time-date, 10-digit stamp to a file with 56,000+ records, all in 7 seconds. Meaning, 8000+ records got the same time-date stam, except for an appendage of 3-character alpha (aaa - zzz). This being said, is there a similar problem that I could expect from using the Unix Epoch 10-digit stamp, where I want this to be a "unique" identifier? In other words, if multiple processes, per second, ask for Unix Epoch, will they all get the same response? I hope I have explained my problem clearly. Thanks, in advance, for any responses. Jim Cronin I have a question for anyone, regarding the use of Unix Epoc time as a “guaranteed, unique” record ID. I manage a system in a non-Unix environment that utilizes a similar “time stamp”, using an internal 5-digit date, concatenated with seconds-since-midnight. If I wish to use this stamp as a record ID, I have to expect that other requests are being made for the time-date stamp, and may find an alpha character appended (1411035660a). I have actual examples of this scenario where the letter “p” has been appended, meaning 15 other applications were asking for time-date stamp, simultaneously. At the speed of today’s processors, this can be expected. In another test, I wrote the time-date, 10-digit stamp to a file with 56,000+ records, all in 7 seconds. Meaning, 8000+ records got the same time-date stam, except for an appendage of 3-character alpha (aaa – zzz). This being said, is there a similar problem that I could expect from using the Unix Epoch 10-digit stamp, where I want this to be a “unique” identifier? In other words, if multiple processes, per second, ask for Unix Epoch, will they all get the same response?

I hope I have explained my problem clearly. Thanks, in advance, for any responses.

Jim Cronin

]]>