python - How to build custom pandas.tseries.offsets class? -


i want find way build custom pandas.tseries.offsets class @ 1 second frequency trading hours. main requirement here time offset object smart enough know next second of '2015-06-18 16:00:00' '2015-06-19 09:30:00 or 09:30:01', , time delta computed these 2 time stamps 1s (custom offset 1s, similar bday(1) business day frequency) instead of duration of closing hours.

the reason when plotting pd.series intraday data across few trading days, see simulated example below, there lots of 'step line' (linear interpolation) between close , next day open prices represent time duration of closing hours. there way rid of this? i've @ source codes of pandas.tseries.offsets , find pd.tseries.offsets.businesshour , pd.tseries.offsets.businessmixin may don't know how use those. 

import pandas pd import numpy np pandas.tseries.holiday import usfederalholidaycalendar pandas.tseries.offsets import custombusinessday  # set 'constant' object shared codes in script bday_us = custombusinessday(calender=usfederalholidaycalendar()) sample_freq = '5min' dates = pd.date_range(start='2015-01-01', end='2015-01-31', freq=bday_us).date # exculde 09:30:00 included in first time bucket times = pd.date_range(start='09:30:00', end='16:00:00', freq=sample_freq).time[1:] time_stamps = [dt.datetime.combine(date, time) date in dates time in times] s = pd.series(np.random.randn(len(time_stamps)).cumsum() + 100, index=time_stamps)  s.plot()

enter image description here

another way can think of partially fix problem first reset_index() default consecutive integer index each row, , calculating difference between consecutive integer index time(in seconds) elapsed. plotting integer index x-axis , relabel them appropriate time labels. show me how matplotlib well?

thanks jeff's comments. check on-line docs businesshour() , find may useful in case. follow-up question(s): businesshour in hour frequency, there way make @ 1s frequency? also, how combine custombusinessday object?

to use businesshour()

from pandas.tseries.offsets import * bhour = businesshour(start='09:30', end='16:00') time = pd.timestamp('2015-06-18 15:00:00') print(time) 2015-06-18 15:00:00 # hourly increment works nicely print(time + bhour * 1) 2015-06-19 09:30:00 # not @ minute or second frequency print(time + minute(61)) 2015-06-18 16:01:00 print(time + second(60*60 + 1)) 2015-06-18 16:00:01

many many thanks, , highly appreciated.

as mention in comment, potentially have 2 different problems

  1. you need able plot business times timeseries without long linear interpolations.
  2. you need object can datetime arithmetic (in seconds) ignoring non-business times

i've given solution account 1 seems immediate issue. if need 2, or both - let know in comments:

1. plot points business days adjacent

most graphs in matplotlib can have index formatters applied axes via ticker api. i'll adapt this example suit case 

import pandas pd import numpy np pandas.tseries.holiday import usfederalholidaycalendar pandas.tseries.offsets import custombusinessday import datetime dt import matplotlib.pyplot plt import matplotlib.ticker ticker  # set 'constant' object shared codes in script bday_us = custombusinessday(calender=usfederalholidaycalendar()) sample_freq = '5min' dates = pd.date_range(start='2015-01-01', end='2015-01-31', freq=bday_us).date # exculde 09:30:00 included in first time bucket times = pd.date_range(start='09:30:00', end='16:00:00', freq=sample_freq).time[1:] time_stamps = [dt.datetime.combine(date, time) date in dates time in times] s = pd.series(np.random.randn(len(time_stamps)).cumsum() + 100, index=time_stamps)  data_length = len(s) s.index.name = 'date_time_index' s.name='stock_price' s_new = s.reset_index()  ax = s_new.plot(y='stock_price') #plot data against new linearised index...  def format_date(x,pos=none):     thisind = np.clip(int(x+0.5), 0, data_length-1)     return s_new.date_time_index[thisind].strftime('%y-%m-%d %h:%m:%s')  ax.xaxis.set_major_formatter(ticker.funcformatter(format_date))  fig = plt.gcf() fig.autofmt_xdate()  plt.show()

this gives output follows, first @ zoomed out natural scale, second zoomed in can see transition between friday 16:00 , monday 09:00

zoomed out time series

zoomed in time series on weekend


Comments

Post a Comment

Popular posts from this blog

powershell Start-Process exit code -1073741502 when used with Credential from a windows service environment -

i'm running strange behavior powershell start-process call. here call: $process = start-process ` "c:\somepath\mybinary.exe" ` -passthru ` -credential $defaultcredential ` -wait ` -workingdirectory "c:\somepath" ` -loaduserprofile if ($process.exitcode -ne 0) { #do } this call return exit code - 1073741502 . after quick search, exit code seems related generic error when program not load required dll (aka. status_dll_init_failed ). when run without -credential $credential program runs correctly. in order isolate problem, manually launched some.exe in prompt target credential , runs smoothly. so problem seems come way start-process cmdlet launch process. i found potential solutions problem tried apply no luck : link , link . would have idea of what's going on here ? edit 1: run proc mon monitoring program activities when launched directly or via powershell script. problem seems occur when loading kernelbase
Read more

twig - Using Twigbridge in a Laravel 5.1 Package -

i'm trying create laravel 5.1 package uses twigbridge instead of .php views. can normal .php views display fine, when try make use of .twig templates i'm met error. possible use twigbridge templates when creating views laravel package or pretty stuck blade , php ? edit 1: loading twig template without extending works. it's when start using {% extends 'file' %} error comes out. i error error loading /virtual/laravel/packages/username/mypackage/src/views/sample.twig: template "templates.master" not defined () in "/virtual/laravel/packages/username/mypackage/src/views/sample.twig" @ line 1. my folders username |_ mypackage |_ src |_ controllers - samplecontroller.php |_ views |_ templates - master.twig - sample.twig twig template sample.twig {% extends 'templates.master' %} {% block content %} {{ parent()
Read more

c# - LINQ join Entities from HashSet's, Join vs Dictionary vs HashSet performance -

i have hashset's each stores t, have written test application compares different relation algorithms can think of, i'm not pleased results getting. does there exist more efficient ways of achieving object relations, once testing? using system; using system.collections.generic; using system.linq; using system.text; using system.threading.tasks; using system.diagnostics; namespace linqtests { class program { static void main(string[] args) { hashset<user> usertable = new hashset<user>(); hashset<userproperty> userpropertytable = new hashset<userproperty>(); #region lets create dummy data. console.writeline("please wait while creating dummy data, can take while..."); console.writeline(""); int rows = 1000000; for(int x = 0; x < rows; x++) { random rnd = new random(); // ad
Read more