Using the default delta time field for Django, but as most know it returns horrifically. So I have developed my own function, convert_timedelta(duration). It computes days, hours, minutes and seconds, just like you might find on this website. It works like a charm on a single time ( I.e. when I was calculating an average time for one specific column ) .. However now I'm modifying how this works so it returns multiple times for a column... Separating times returned grouped by their ID, rather than just having one set time returned. So now I have multiple times returned, which works fine with the default formatting. But when I apply it to my function it doesn't like it and no particular informative errors are alerted.
def convert_timedelta(duration):
days, seconds = duration.days, duration.seconds
hours = days * 24 + seconds // 3600
minutes = (seconds % 3600) // 60
seconds = (seconds % 60)
Related
I would like to specify a date in the condition of the function "if", how can I do that?
For example, is it possible to set a condition such as "if the closing price rises by 10% within a week"?
In order to get the close price one week ago, you'll need to reference a bar in history. You can get the number of bars you need to reference in history using this code:
bars_back = (time - (time - 604800000)) / (1000 * timeframe.in_seconds(timeframe.period))
From there it's just comparing the condition:
if close / close[bars_back] > 1.1
// do something
Example code:
//#version=5
indicator("My script", overlay = true)
bars_back = (time - (time - 604800000)) / (1000 * timeframe.in_seconds(timeframe.period))
cond = close / close[bars_back] > 1.1
plotshape(cond)
I need to calculate the total amount of time for a certain number of tasks to be completed. Details:
5 tasks total. Time estimates (in seconds) for each: [30, 10, 15, 20, 25]
Concurrency: 3 tasks at a time
How can I calculate the total time it will take to process all tasks, given the concurrency? I know it will take at least as long as the longest task (25 seconds), but is there a formula/method to calculate a rough total estimate, that will scale with more tasks added?
If you don't mind making some approximations it could be quite simple. If the tasks take roughly the same time to complete, you could use the average of the tasks duration as a basis (here, 20 seconds).
Assuming that the system is always full of tasks, that task duration is small enough, that there are many tasks and that concurrency level is high enough, then:
estimated_duration = average_task_duration * nb_tasks / nb_workers
Where nb_workers is the number of concurrent threads.
Here is some Python code that shows the idea:
from random import random
from time import sleep, monotonic
from concurrent.futures import ThreadPoolExecutor
def task(i: int, duration: float):
sleep(duration)
def main():
nb_tasks = 20
nb_workers = 3
average_task_duration = 2.0
expected_duration = nb_tasks * average_task_duration / nb_workers
durations = [average_task_duration + (random() - 0.5) for _ in range(nb_tasks)]
print(f"Starting work... Expected duration: {expected_duration:.2f} s")
start = monotonic()
with ThreadPoolExecutor(max_workers=nb_workers) as executor:
for i, d in enumerate(durations):
executor.submit(task, i, d)
stop = monotonic()
print(f"Elapsed: {(stop - start):.2f} s")
if __name__ == "__main__":
main()
If these hypotheses cannot hold in your case, then you'd better use a Bin Packing algorithm as Jerôme suggested.
I'm working with call center data and looking to calculate the average ring time of calls while removing the highest 20% of ring times. I assume I'll need to use PERCENTILEX.EXC embedded somewhere in AVERAGE, but I'm not quite sure where, or if I'm totally off base. 2 other caveats on this are that there are calls answered immediately (queue time = 0) which have to be counted in the average time and only data where the disposition column = Handled are used.
Example:
The Aborted and Abandoned call would be filtered out. Of the remaining calls, the top 20% of queue times (the 14,9, 6, and one of the 5s) would be eliminated and the average would be 3 seconds.
Appreciate any help on this!
I would do it like this:
VAR totalRows = COUNTROWS(FILTER(table, table[disposition] = "Handled"))
VAR bottomN = ROUNDDOWN(totalRows * .8, 0)
RETURN AVERAGEX(TOPN(bottomN, FILTER(table, table[disposition] = "Handled"), table[queue time], ASC),table[queue time])
I am developing a python application using Python2.7.
That application is using pyserial library to read some data bytes from serial port.
Using while loop to read 1 byte of data in each iteration. In each iteration
I have to measure execution time between statements if it is less than 10 ms wait for it to reach 10ms before starting next iteration. there are two questions here as follows:
Time measurement
What would be the best way to measure time between python statements to the accuracy (1ms or 2ms difference acceptable) of milliseconds.
Time delay
How can i use that measured time for delay in order to wait till total time is 10ms (total time = 10ms = code execution time+delay)
I have tried time library but it does not give a good resolution when in millisecond some time it does not give anything when small time duration.
for example:
import time
while uart.is_open:
uart.read()
start = time.time()
#user code will go here
end = time.time()
execution_time=end - start
if execution_time< 10ms:
remaining_time = 10ms - execution_time
delay(remaining_time)
You can get a string that is minutes:seconds:microseconds using datetime like so:
import datetime
string = datetime.datetime.now().strftime("%M:%S:%f")
And then turn it into a number to make comparison more handy:
m, s, u = string.split(":")
time_us = float(m)/60e6 + float(s)/1e6 + float(u)
In your example, that would look like this:
import datetime
def time_us_now():
h, m, s, u = (float(string) for string in datetime.datetime.now().strftime("%H:%M:%S:%f").split(":"))
return h/3600e6 + m/60e6 + s/1e6 + u
start = time_us_now()
# User code will go here.
end = time_us_now()
execution_time = end - start
Solution:
I have managed to get upto 1ms resolution by doing two things.
Changed (increased) the baud rate to 115200 from 19200.
Used time.clock() rather than time.time() as it has more accuracy and resolution.
The release notes for UniVerse version 11.2.4 mention local time zone configuration, but it is in the context of auditing. This is the quote:
Local time zone configuration
Prior to UniVerse 11.2.4, the date and time data stored in the audit log records was based on UTC only.
Beginning at UniVerse 11.2.4, UniVerse adds the date and time data
based on local timezone to audit log records. The data is stored in
location 19 for each record. The dictionary name for this data field
is TZINFO. For more information, see UniVerse Security Features.
Since UniVerse seems capable of working with time zones natively, does this mean there might be a way to easily generate UTC-formatted date/time stamps from my EST/EDT values?
I am sending data to a system that wants dates formatted in ISO-8601 Date/Time format yyyy-MMddTHH:mm:ssZ, like 2015-06-02T15:55:22Z, with the time zone and Daylight Saving Time offsets accounted for.
I dug through the Security Features guide, and found this:
UniVerse also adds a globally cataloged program to help users to
obtain date and time information from the audit log (which is called
by the above two I-descriptor fields):
SUBROUTINE GETLOCALTIME
(
RESULT ;* OUT: output
TZOFF ;* IN: time zone offset
DATE ;* IN: UTC date
TIME ;* IN: UTC time
OP ;* IN: operation
;* 1: get local date
;* 2: get local time
;* 3: get local timezone id
;* 4: get local timezone daylight saving flag
)
(Since I'm not using the auditing capabilities of UniVerse, I don't think I can do much with this, nor could I locate the subroutine.)
I have also played with the popular(?) DATE.UTILITY program from PickWiki, but its calculation of Daylight Saving Time start/end dates seem off. I will save those issues for another question.
This is getting long-winded but I'm hoping someone can point me in the right direction if there's a new OCONV() parameter or something I could use.
Just in case it matters, I'm running on Windows Server 2008 R2.
Thanks!
Time is a complicated thing. Socially we have accepted that it not only acceptable to alter it 2 times a year, we have mandated it! This is all well and good for us meat machines who only want to understand time when it is convenient for us however it does cause us to get grumpy when out reporting "looks funny".
The solution to your problem is not exceptionally easy, especially if you are working with already recorded dates. Dates and times in Universe are generally recorded based on local system time. If this is something that you are trying to do going forward you have to note what the offset is at the time of the transaction or simply stamp things SYSTEM(99), which complicated pretty much all other reporting you will need to do. Either way, this is a complicated matter and it still likely to be somewhat imperfect.
Here is a little something that might help you if you are the one in charge of recording dates, going forward.
SECONDS.SINCE.GMT.01.01.1970 = SYSTEM(99)
CRT SECONDS.SINCE.GMT.01.01.1970:" Seconds since GMT Epoch Began"
NUMBER.OF.DAYS.SINCE.01.01.1970 = DATE() -732
;* Day 0 in Pick is 12/31/1967 because Dick Pick so we subtract 732 from the pick date
SECONDS.SINCE.MIDNIGHT.LOCAL= TIME()
SECS.PER.DAY = 24 * 60 * 60
LOCAL.SECONDS.SINCE.GMT.01.01.1970 = NUMBER.OF.DAYS.SINCE.01.01.1970 * SECS.PER.DAY + FIELD(SECONDS.SINCE.MIDNIGHT.LOCAL,".",1)
;*I drop the precision
CRT LOCAL.SECONDS.SINCE.GMT.01.01.1970: " Seconds since 01/01/1970 in local time"
OFFSET = (LOCAL.SECONDS.SINCE.GMT.01.01.1970 - SECONDS.SINCE.GMT.01.01.1970)
CRT "CURRENT.OFFSET IS ":INT((OFFSET / 60 )/ 60)
END
Which outputs the following on my system which is currently PDT (even though OCONV(DATE(),'DZ') reports it as PST.
1434472817 Seconds since GMT Epoch Began
1434447617 Seconds since 01/01/1970 in local time
CURRENT.OFFSET IS -7
Hopefully you have found this helpful.
Thanks for the clues. Here's my implementation:
SUBROUTINE FORMAT.ISO.8601 (IDATE, ITIME, RESULT, ERR.TEXT)
* Don't step on the caller's variables.
IN.DATE = IDATE
IN.TIME = ITIME
* Initialize the outbound variable.
RESULT = ''
IF NOT(NUM(IN.DATE)) THEN
ERR.TEXT = 'Non-numeric internal date ' : DQUOTE(IN.DATE) : ' when numeric required.'
RETURN
END
IF NOT(NUM(IN.DATE)) THEN
ERR.TEXT = 'Non-numeric internal time ' : DQUOTE(IN.TIME) : ' when numeric required.'
RETURN
END
* SYSTEM(99) is based on 1/1/1970.
SECONDS.SINCE.GMT.01.01.1970 = SYSTEM(99)
* Day 0 in Pick is 12/31/1967
* Subtract 732 to equalize the starting dates.
NUMBER.OF.DAYS.SINCE.01.01.1970 = DATE() - 732
SECONDS.SINCE.MIDNIGHT.LOCAL= TIME()
SECS.PER.DAY = 24 * 60 * 60
LOCAL.SECONDS.SINCE.GMT.01.01.1970 = NUMBER.OF.DAYS.SINCE.01.01.1970 * SECS.PER.DAY + FIELD(SECONDS.SINCE.MIDNIGHT.LOCAL,".",1)
OFFSET = LOCAL.SECONDS.SINCE.GMT.01.01.1970 - SECONDS.SINCE.GMT.01.01.1970
OFFSET = INT((OFFSET / 60 )/ 60)
OTIME = OCONV(IN.TIME, 'MTS')
IF OTIME = '' THEN
ERR.TEXT = 'Bad internal time ' : DQUOTE(IN.TIME) : '.'
RETURN
END
HOURS = FIELD(OTIME, ':', 1)
MINUTES = FIELD(OTIME, ':', 2)
SECONDS = FIELD(OTIME, ':', 3)
HOURS -= OFFSET
IF HOURS >= 24 THEN
IN.DATE += 1
HOURS = HOURS - 24
END
HOURS = HOURS 'R%2'
ODATE = OCONV(IN.DATE, 'D4/')
IF ODATE = '' THEN
ERR.TEXT = 'Bad internal date ' : DQUOTE(IN.DATE) : '.'
RETURN
END
DMONTH = FIELD(ODATE, '/', 1)
DDAY = FIELD(ODATE, '/',2)
DYEAR = FIELD(ODATE, '/',3)
RESULT = DYEAR : '-' : DMONTH : '-' : DDAY : 'T' : HOURS : ':' : MINUTES : ':' : SECONDS : 'Z'
RETURN
END
Here's my test harness:
CRT 'Testing right now.'
IDATE = DATE()
ITIME = TIME()
CALL FORMAT.ISO.8601 (IDATE, ITIME, RESULT, ERR.TEXT)
IF ERR.TEXT THEN
CRT 'ERR.TEXT: ' : ERR.TEXT
END ELSE
CRT 'RESULT: ' : RESULT
END
CRT
CRT 'Testing an hour ago.'
IDATE = DATE()
ITIME = TIME()
ITIME = ITIME - (60*60)
IF ITIME < 0 THEN
ITIME += (24*60*60)
IDATE -= 1
END
CALL FORMAT.ISO.8601 (IDATE, ITIME, RESULT, ERR.TEXT)
IF ERR.TEXT THEN
CRT 'ERR.TEXT: ' : ERR.TEXT
END ELSE
CRT 'RESULT: ' : RESULT
END
CRT
CRT 'Testing an hour from now.'
IDATE = DATE()
ITIME = TIME()
ITIME = ITIME + (60*60)
IF ITIME > (24*60*60) THEN
ITIME -= (24*60*60)
IDATE += 1
END
CALL FORMAT.ISO.8601 (IDATE, ITIME, RESULT, ERR.TEXT)
IF ERR.TEXT THEN
CRT 'ERR.TEXT: ' : ERR.TEXT
END ELSE
CRT 'RESULT: ' : RESULT
END
END
Here's my test run:
>T$FORMAT.ISO.8601
Testing right now.
RESULT: 2017-03-29T00:47:22Z
Testing an hour ago.
RESULT: 2017-03-28T23:47:22Z
Testing an hour from now.
RESULT: 2017-03-29T01:47:22Z