*This macro takes a begin and end datetime in quotes, and then calculates steps.
*Useful for generating aggregate statistics.


DEFINE !TimeExpand (Begin = !TOKENS(1)
                   /End = !TOKENS(1)
                   /Step = !DEFAULT("60*60*24") !TOKENS(1)
                   /Var = !DEFAULT (DT) !TOKENS(1)
                   /Dataset = !DEFAULT ("TimeExpan") !TOKENS(1) )
INPUT PROGRAM.
COMPUTE #Begin = !UNQUOTE(!Begin).
COMPUTE #End = !UNQUOTE(!End).
COMPUTE #Step = !UNQUOTE(!Step).
LOOP #i = #Begin TO #End BY #Step.
  COMPUTE !Var = #i.
  END CASE.
END LOOP.
END FILE.
END INPUT PROGRAM.
DATASET NAME !UNQUOTE(!Dataset).
FORMATS !Var (DATETIME17).
EXECUTE.
!ENDDEFINE.

*Example with defaults.
*TimeExpand Begin = "DATE.MDY(01,01,2013)" End = "DATE.MDY(01,01,2014)".

*15 minute example.
*DATASET CLOSE ALL.
*TimeExpand Begin = "DATE.MDY(01,01,2013)" End = "DATE.MDY(01,01,2014) - TIME.HMS(0,10)" Step = "15*60" Dataset = "Time15min".

*Week from a begin date example.
*DATASET CLOSE ALL.
*TimeExpand Begin = "DATE.MDY(01,01,2013)" End = "DATE.MDY(01,01,2014)" Step = "60*60*24*7" Dataset = "WeekBin".

*Backwards!.
*DATASET CLOSE ALL.
*TimeExpand Begin = "DATE.MDY(01,01,2014)" End = "DATE.MDY(01,01,2013)" Step = "-60*60*24*7" Dataset = "WeekBinBack".

*1 Minute example.
*DATASET CLOSE ALL.
*TimeExpand Begin = "DATE.MDY(01,01,2013)" End = "DATE.MDY(01,01,2014) - TIME.HMS(0,0,1)" Step = "60" Dataset = "Time1min".

*You can use current date-time.
*DATASET CLOSE ALL.
*TimeExpand Begin = "DATE.MDY(01,01,2013)" End = "XDATE.DATE($TIME)".


*Valid types are years, quarters, months, weeks, days, hours, minutes, seconds.
DEFINE !DateExpand (Begin = !TOKENS(1)
                   /End = !TOKENS(1)
                   /Type = !TOKENS(1)
                   /Var = !DEFAULT (DT) !TOKENS(1)
                   /Dataset = !DEFAULT ("DateExpan") !TOKENS(1) )
INPUT PROGRAM.
COMPUTE #B = !UNQUOTE(!Begin).
COMPUTE #E = !UNQUOTE(!End).
COMPUTE #Int = DATEDIFF(#E,#B,!Type).
LOOP #i = 0 TO #Int.
  COMPUTE !Var = DATESUM(#B,#i,!Type).
  END CASE.
END LOOP.
END FILE.
END INPUT PROGRAM.
DATASET NAME !UNQUOTE(!Dataset).
FORMATS !Var (DATETIME17).
!ENDDEFINE.

*DATASET CLOSE ALL.
*DateExpand Begin = "DATE.MDY(01,01,2013)" End = "DATE.MDY(01,01,2014)" Type = "DAYS".

*DATASET CLOSE ALL.
*DateExpand Begin = "DATE.MDY(01,01,2013)" End = "DATE.MDY(01,01,2014)" Type = "MONTHS" Dataset = "DateMon".

*DATASET CLOSE ALL.
*DateExpand Begin = "DATE.MDY(01,01,2013)" End = "$TIME" Type = "MINUTES" Dataset = "DateMin".


*These macros are the same except fill in a panel dataset of given ids.

DEFINE !PanelFillTime (OrigData = !TOKENS(1)
                    /Id = !TOKENS(1) 
                    /DateVar = !TOKENS(1)
                    /Begin = !TOKENS(1)
                    /End = !TOKENS(1)
                    /Step = !DEFAULT("60*60*24") !TOKENS(1)                        
                    /NewData = !DEFAULT ("TimeExpan") !TOKENS(1) )
*Making dataset with just the IDs.
DATASET ACTIVATE !OrigData.
SORT CASES BY !DateVar !Id.
DATASET DECLARE Ids.
AGGREGATE OUTFILE='Ids' /BREAK=!Id /!DateVar = FIRST(!DateVar).
DATASET ACTIVATE Ids.
SELECT IF NOT MISSING(!Id).
COMPUTE Const = 1.
CASESTOVARS /ID = Const /DROP !DateVar.
NUMERIC @End.
*Making expanded time set.
INPUT PROGRAM.
COMPUTE #Begin = !UNQUOTE(!Begin).
COMPUTE #End = !UNQUOTE(!End).
COMPUTE #Step = !UNQUOTE(!Step).
LOOP #i = #Begin TO #End BY #Step.
  COMPUTE !DateVar = #i.
  END CASE.
END LOOP.
END FILE.
END INPUT PROGRAM.
DATASET NAME !UNQUOTE(!NewData).
DATASET ACTIVATE !UNQUOTE(!NewData).
FORMATS !DateVar (DATETIME17).
COMPUTE Const = 1.
*Merging in IDs.
MATCH FILES FILE = *
  /TABLE = 'Ids'
  /BY Const.
DATASET CLOSE Ids.
VARSTOCASES /MAKE !Id FROM !CONCAT(!Id,".1") TO @End /DROP Const.
MATCH FILES FILE = *
  /TABLE = !QUOTE(!OrigData)
  /BY !DateVar !Id.
!ENDDEFINE.

*Testing it out.
*DATA LIST FREE / MyId (F1.0) DateX (ADATE8) DataEx.
*BEGIN DATA
*1 1/1/2010 1
*1 1/3/2010 2
*1 1/5/2010 3
*2 1/2/2010 4
*2 1/3/2010 5
*2 1/5/2010 6
*END DATA.
*DATASET NAME AggPan.

*PanelFillTime OrigData = AggPan Id = MyId DateVar = DateX Begin = "DATE.MDY(1,1,2010)" End = "DATE.MDY(1,6,2010)".

*If you want the missing dates to be zero, simple as.
*RECODE DataEx (SYSMIS=0)(ELSE=COPY).
*If you want the missing dates to be the last value, use.
*SORT CASES BY MyId DateX.
*IF MISSING(DataEx) DataEx = LAG(DataEx).
*EXECUTE.

*Valid types are years, quarters, months, weeks, days, hours, minutes, seconds.
DEFINE !PanelFillDate (OrigData = !TOKENS(1)
                      /Id = !TOKENS(1) 
                      /DateVar = !TOKENS(1)
                      /Begin = !TOKENS(1)
                      /End = !TOKENS(1)
                      /Type = !TOKENS(1)                        
                      /NewData = !DEFAULT ("DateExpan") !TOKENS(1) )
*Making dataset with just the IDs.
DATASET ACTIVATE !OrigData.
SORT CASES BY !DateVar !Id.
DATASET DECLARE Ids.
AGGREGATE OUTFILE='Ids' /BREAK=!Id /!DateVar = FIRST(!DateVar).
DATASET ACTIVATE Ids.
SELECT IF NOT MISSING(!Id).
COMPUTE Const = 1.
CASESTOVARS /ID = Const /DROP !DateVar.
NUMERIC @End.
*Making expanded time set.
INPUT PROGRAM.
COMPUTE #B = !UNQUOTE(!Begin).
COMPUTE #E = !UNQUOTE(!End).
COMPUTE #Int = DATEDIFF(#E,#B,!Type).
LOOP #i = 0 TO #Int.
  COMPUTE !DateVar = DATESUM(#B,#i,!Type).
  END CASE.
END LOOP.
END FILE.
END INPUT PROGRAM.
DATASET NAME !UNQUOTE(!NewData).
DATASET ACTIVATE !UNQUOTE(!NewData).
FORMATS !DateVar (DATETIME17).
COMPUTE Const = 1.
*Merging in IDs.
MATCH FILES FILE = *
  /TABLE = 'Ids'
  /BY Const.
DATASET CLOSE Ids.
VARSTOCASES /MAKE !Id FROM !CONCAT(!Id,".1") TO @End /DROP Const.
MATCH FILES FILE = *
  /TABLE = !QUOTE(!OrigData)
  /BY !DateVar !Id.
!ENDDEFINE.

*Testing it out.
*DATA LIST FREE / MyId (F1.0) DateX (ADATE8) DataEx.
*BEGIN DATA 
*1 1/1/2010 1
*1 1/3/2010 2
*1 1/5/2010 3
*2 1/2/2010 4
*2 1/3/2010 5
*2 1/5/2010 6
*END DATA.
*DATASET NAME AggPan.

*PanelFillDate OrigData = AggPan Id = MyId DateVar = DateX Begin = "DATE.MDY(1,1,2010)" End = "DATE.MDY(1,6,2010)" Type = "HOURS".