|
[^]
Caveat Emptor.
"Progress doesn't come from early risers – progress is made by lazy men looking for easier ways to do things." Lazarus Long
|
|
|
|
|
That's your tech gap manifest right there.
|
|
|
|
|
That is incorrect. CMOS circuits still consume current, even when they are not switching. It's commonly referred to as leakage, and is well understood. There are many sources of leakage in CMOS circuits, but the most common is subthreshold leakage. Generally speaking, this leakage gets worse as technology gets smaller, but - in recent years - has been mitigated against with fin-FET technologies.
|
|
|
|
|
I stand corrected. Thank you.
|
|
|
|
|
Ok, I swept leakage under the rug, but the lion's share goes to switching and multiplies with the frequency and the number of transistors. And that ends up entirely as heat we must get rid of.
I am endeavoring, ma'am, to construct a mnemonic memory circuit using stone knives and bearskins.
|
|
|
|
|
Quote: And that ends up entirely as heat and cat videos we must get rid of
FTFY
|
|
|
|
|
...exploring the msdb.dbo.sysschedules table in Sql Server, and one of the columns is called active_end_time of type int.
When I saw the definition, I assumed that it would represent the number of seconds in the day (from midnight). I was wrong (and surprised, bewildered, and downright annoyed) when presented with the reality of the situation.
The value returned was 235959 which far exceeds the number of seconds in a day (86400). In reality, it is a numeric representation of "23:59:59". WTF Microsoft!!!! What rocket scientist decided that this was a valid way to represent a time-of-freakin-day?!
Even worse, they don't use two-digit hours, so "30000" represents "3 am".
Also, "0" indicates midnight, so how do they represent 1 second past midnight? (I don't know because none of existing the job schedules have a time like that.)
EDIT =================
Out of curiosity, I added a job with a schedule that starts at 00:00:01, and the value stored in table is "1".
".45 ACP - because shooting twice is just silly" - JSOP, 2010
- You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
- When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
modified 4-Oct-17 13:02pm.
|
|
|
|
|
John Simmons / outlaw programmer wrote: I assumed that it would represent the number of seconds in the day (from midnight) I would have, too. It's odd they chose the format they have, but at least it's ordinal and documented[^].
/ravi
|
|
|
|
|
I created an extension method to parse it...
public static class ExtendTimeSpan
{
public static TimeSpan ParseFromFormattedInt(int value)
{
return new TimeSpan(0).ParseFromFormattedInt(value);
}
public static TimeSpan ParseFromFormattedInt(this TimeSpan span, int value)
{
TimeSpan result = new TimeSpan(0);
string timeString = value.ToString();
string padding = new String('0', 6 - timeString.Length);
timeString = ((timeString.Length < 6) ? string.Concat(padding, timeString) : timeString).Insert(4, ":").Insert(2,":");
try
{
result = TimeSpan.ParseExact(timeString, "g", CultureInfo.CurrentCulture);
}
catch (Exception)
{
}
return result;
}
}
".45 ACP - because shooting twice is just silly" - JSOP, 2010
- You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
- When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
|
|
|
|
|
/ravi
|
|
|
|
|
No need to mess around with strings:
public static class ExtendTimeSpan
{
public static TimeSpan ParseFromFormattedInt(int value)
{
int hours = Math.DivRem(value, 10000, out value);
int minutes = Math.DivRem(value, 100, out value);
int seconds = value + minutes * 60 + hours * 3600;
return TimeSpan.FromSeconds(seconds);
}
public static TimeSpan ParseFromFormattedInt(this TimeSpan span, int value)
{
return ParseFromFormattedInt(value);
}
}
Or, in SQL (2012 or later):
SELECT
TIMEFROMPARTS(active_end_time / 10000, (active_end_time / 100) % 100, active_end_time % 100, 0, 0)
FROM
msdb.dbo.sysschedules
;
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Richard Deeming wrote: Or, in SQL (2012 or later):
This is 2008R2... :/
".45 ACP - because shooting twice is just silly" - JSOP, 2010
- You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
- When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
|
|
|
|
|
I knew a math way was possible. This can also be extended to the dates that are stored in the same fashion:
Given an int value of 99991231:
int year = Math.DivRem(value, 10000, out value);
int month = Math.DivRem(value, 100, out value);
int day = value;
return new DateTime(year, month, day);
would yield a datetime of 12/31/9999.
".45 ACP - because shooting twice is just silly" - JSOP, 2010
- You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
- When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
modified 5-Oct-17 8:53am.
|
|
|
|
|
As if integer time wasn't bad enough!
We do a lot of work with a Norwegian ERP system that stores dates like that. The UI validates the day and month, but doesn't validate the year. We regularly have to correct user input like 200171005 or 2011130 .
It's also allergic to Null in the database. If a date hasn't been entered, it's stored as 0 .
Fun times.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
If 235959 represents 23:59:59 and the type is 'int', how would they have a two digit hour for 3 AM?
It seem reasonable that any time represented as a 6 character string without ":" delimiters, when converted to a 'int' would give the appropriate response.
So.. a time of 00:00:01, when stripped of delimiters would give 000001 and converted to an 'int' would give... 1.
So.. outside of the odd data type, what's the issue?
|
|
|
|
|
The issue is that it's an absurd way to represent the time.
They're using an int data type, so the number of seconds past midnight would have made more sense, because it requires the same amount of storage space to store 1 as it does to store 1000.
This means that in order to determine "the time", I had to write a method to parse the value instead of using
TimeSpan span = TimeSpan.FromSeconds(1);
".45 ACP - because shooting twice is just silly" - JSOP, 2010
- You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
- When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
|
|
|
|
|
I feel your pain. I have spent a considerable amount of time trying to unravel a MySql database that someone else built. Various dates were stored as integers, decimal, string, in fact anything except the obvious DATE!
How do you know if a date value of 1216 represents 1st February 2016, 12th January 2006 or what?
Grrr!
We're philosophical about power outages here. A.C. come, A.C. go.
|
|
|
|
|
It's silly, but I see the logic: it's human readable, as well as easily calculated for matches. "Number of seconds since midnight" is a better solution except for the human readable bit: when is 52642? Can you tell exactly and easily without a calculator? Or is it easier to just look at 143722 and know immediately when it is?
The silliness if you think about it is having 60 seconds to the minute, 60 minutes to the hour, 24 hours to the day - instead of using a minute that was about 50% longer than the current one, made up of 100 (slightly longer) seconds, and having 1000 new-minutes in a day.
We could call it "Stardate" and annoy rabid Trekkies.
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
OriginalGriff wrote: It's silly, but I see the logic
The logic is that you don't invent yet another representation for time values...
Who felt this was necessary?
|
|
|
|
|
I cannot tell you who thought it was necessary, but it is very useful for report builders. You have a stored procedure running daily and you concatenate the numeric date on the end without needing conversions:
MyReportOnFinances_20171005.csv
|
|
|
|
|
OriginalGriff wrote: "Number of seconds since midnight" is a better solution except for the human readable bit:
It's a numeric value in a database - it doesn't have to be human readable as the actual time. It requires parsing to be used, especially if you want to perform math with it (the SQL required to parse this would be a mild nightmare), and almost always when you simply want to display it. I stand by my original claim that it's absurd.
".45 ACP - because shooting twice is just silly" - JSOP, 2010
- You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
- When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
|
|
|
|
|
Strikes me that this might be a hangover from the days when memory/disk was in short supply and storing an INT took 16 bits, compared with a 6 char string which took 48. If not directly so, then because the programmer who wrote the code had a history of working on such systems...
(First release of SQL Server was 16bit, for OS/2 in 1989 I believe)
There was a time when memory was severely constrained, not to mention thousands and thousands of times more expensive than it is now, and likewise processor/disk speed was much slower, so this kind of approach was needed.
It's only recently that we have got used to a web page running in a browser using more computing power than an entire organisation once possessed to simply display an annoying pop-up ad with sound and video.
8)
|
|
|
|
|
But no of secs in a day = 86400 (17 bits if signed int, 16 bits if unsigned int)
However, 235959 needs 18 bits if signed int or 17 bits if unsigned int
|
|
|
|
|
You are right, of course. I don't know how things were/are stored internally in SQL Server, but even in 16 bit stuff you could usually store 32bit INTs, which is still smaller than the 48bits needed for "235959". You could usually manipulate DWORDS (32 bits) with single machine instructions, where as to manipulate "235959" would require a whole subroutine. Hence my comment.
These days no-one thinks twice (or even once) about using a few megabytes of inefficient code to splurge something onto your browser window, but I suspect in the highly competitive world of high-end/large/NoSQL database implementations where speed is often of the essence, such apparently strange optimisations are still being used.
|
|
|
|
|
Mike Winiberg wrote: You are right, of course. I don't know how things were/are stored internally in SQL Server, but even in 16 bit stuff you could usually store 32bit INTs, which is still smaller than the 48bits needed for "235959".
You would only need 48bits if you were storing as text. 2,147,483,647 can be stored in a signed 32bit INT.
Michael Martin
Australia
"I controlled my laughter and simple said "No,I am very busy,so I can't write any code for you". The moment they heard this all the smiling face turned into a sad looking face and one of them farted. So I had to leave the place as soon as possible."
- Mr.Prakash One Fine Saturday. 24/04/2004
|
|
|
|
|