BIMETS Time Series Indexing
The package extends the way users can access and modify time series data in order to ease time series analysis and manipulation.
SELECTING BY DATE: users can select a single observation by date by using the syntax ts['Date'] and multiple observations by using ts['StartDate/EndDate'] or ts['StartDate'+(0:n)/f], with f=frequency, n=#observations.
Data modification follows the same syntax: ts['Date']=value, ts['Date/Date']=c(value,value,...,value), etc. Users can also provide the string representing only the year of selection, or the year and the month of selection. For quarterly and monthly time series it is possible to select dates by using instances of class yearmon() and yearqtr() (See examples).
SELECTING BY YEAR-PERIOD: Users can select observations by providing the year and the period requested. Selection and modification of data require the double square bracket syntax,
e.g.
ts[[YEAR,PERIOD]]=value. Users can also assign an array of values to the input time series, starting from the [[YEAR,PERIOD]] provided,
i.e. ts[[YEAR,PERIOD]]=c(value1,value2,...,valueN): in this case the input time series will be eventually extended in order to sequentially insert all values of the provided array (See examples).
SELECTING BY INDEXES: (core R) Users can select observations by simply providing the array of requested indexes,
e.g. ts[c(idx1,idx2,...,idxN)] while reading and ts[c(idx1,idx2,...,idxN)]=c(value1,value2,...,valueN) while writing time series data.
#day and month names can change depending on locale
Sys.setlocale('LC_ALL','C')
Sys.setlocale('LC_TIME','C')
#monthly
#-------------------------------
print('MONTHLY GET by DATE')
n=25
#create ts
ts1=TIMESERIES((0:n),START=c(2000,1),FREQ=12)
print(ts1['2001-01']) #get Jan 2001
print(ts1[as.yearmon('Jan 2001')]) #get Jan 2001
print(ts1['2000-09/2001-01']) #get data from Sep 2000 to Jan 2001
print(ts1['2000-09/']) #get data from Sep 2000
print(ts1['/2001-01']) #get data till Jan 2001
print(ts1['2001']) #gat all data in year 2001
#get 3 consecutive months starting from Jan 2001
print(ts1[as.yearmon('Jan 2001')+ 0:2/12])
print(ts1[c(2,4,5)]) #get observation number 2,4 and 5
print('MONTHLY GET by YEAR-PERIOD')
print(ts1[[2000,5]]) #get year 2000 period 5
#get year 2010 period 1 (out of range)
tryCatch({print(ts1[[2010,1]])},error=function(e){cat(e$message)})
print(ts1[[2002,2]]) #get year 2002 period 2
print('MONTHLY SET by DATE')
ts1['2000-08']=9.9 #assign to Aug 2000
ts1[as.yearmon('Feb 2001')]=8.8 #assign to Feb 2001
#assign 8.8 on Feb 2001 and give warning
ts1[as.yearmon('Feb 2001')]=c(8.8,7.7)
#assign same value to all observation in range Sep 2000 - Jan 2001
ts1['2000-09/2001-01']=11.11
#assign repeatedly the two values to each observation starting from Sep 2001
ts1['2001-09/']=c(1.1,2.2)
print(ts1)
print('MONTHLY SET by YEAR-PERIOD')
ts1[[2000,5]]=NA #set year 2000 period 5
#assign an array starting from year 2002 period 2 (extend time series)
ts1[[2002,2]]=c(-1,-2,-3,-4,-5)
TABIT(ts1)
#quarterly
#-------------------------------
print('QUARTERLY GET by DATE')
#create ts
ts1=TSERIES((0:n),START=c(2000,1),FREQ=4)
print(ts1[as.yearqtr('2001 Q1')]) #get 2001 Q1
print(ts1['2001']) #get all data in year 2001
#get 4 consecutive quarters starting from 2002 Q2
print(ts1[as.yearqtr('2002 Q2')+ 0:3/4])
print(ts1['2003/']) #gat all data from 2003 Q1
print('QUARTERLY GET by YEAR-PERIOD')
print(ts1[[2002,4]]) #get year 2002 period 4
print('QUARTERLY SET by DATE')
ts1[as.yearqtr('2001 Q1')]=7.7 #assign to 2001 Q1
ts1['2002']=NA #assign to all observations of 2002
#assign to 3 quaters starting from 2003 Q2
ts1[as.yearqtr('2003 Q2')+ 0:2/4]=0
ts1['2004/']= -1 #assign to all observations starting from 2004
TABIT(ts1)
print('QUARTERLY SET by YEAR-PERIOD')
ts1[[2005,4]]=c(1,2,3) #assign array starting from year 2005 period 4
TABIT(ts1)
#yearly
#-------------------------------
print('YEARLY GET by DATE')
#create ts
ts1=TSERIES((1:n),START=c(2000,1),FREQ=1)
print(ts1['2002-12-31']) #get 2002 data
print(ts1['2002']) #get 2002 data
print(ts1['2000/2004']) #get data from 2000 to 2004
print(ts1['2005/']) #get data starting from 2005
print('YEARLY GET by YEAR-PERIOD')
print(ts1[[2005,1]]) #get year 2005
#get year 2032 (out of range)
tryCatch({print(ts1[[2032,1]])},error=function(e){cat(e$message)})
print('YEARLY SET by DATE')
ts1['2004']=NA #assign to 2004
ts1['2007/']=0.0 #assign starting from 2007
ts1['2000/2002']= -1 #assign in range 2000/2002
TABIT(ts1)
print('YEARLY SET by YEAR-PERIOD')
ts1[[2005,1]]=NA #assign to 2005
ts1[[2014,1]]= c(-1,-2,-3) #assign array starting from 2014 (extend series)
TABIT(ts1)
#daily
#-------------------------------
print('DAILY GET by DATE')
#create ts
ts1=TSERIES((1:n),START=c(2000,1),FREQ='D')
print(ts1['2000-01-12']) #get Jan 12, 2000 data
print('DAILY GET by YEAR-PERIOD')
print(ts1[[2000,14]]) #get year 2000 period 14
#get year 2032 (out of range)
tryCatch({print(ts1[[2032,1]]) },error=function(e){cat(e$message)})
print('DAILY SET by DATE')
ts1['2000-01-15']=NA #assign to Jan 15, 2000
TABIT(ts1)
print('DAILY SET by YEAR-PERIOD')
ts1[[2000,3]]=NA #assign to Jan 3, 2000
#assign array starting from 2000 period 35 (extend series)
ts1[[2000,35]]= c(-1,-2,-3)
TABIT(ts1)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.