impute zeroes when plotting frequencies with time axis

Quentin — Mon, 13 Aug 2018 17:32:53 GMT

I'm using PROC SGPLOT VLINE statement to make a series plot that plots the frequency of an event over time (by month). Events are sparse, so some months have no events

Given data like below (two failures in January, one failure in February, three failures in March, one failure in June).

data have ;
input FailDate mmddyy10.;
cards ;
1/1/2018
1/2/2018
2/1/2018
3/1/2018
3/2/2018
3/3/2018
6/1/2018
;

I'm plotting like:

proc sgplot data=have;
  xaxis type=time interval=month ;
  vline FailDate / stat=freq markers;
  format FailDate yymon. ;
run ;

The resulting plot shows four points (Jan, Feb, Mar, Jun). I would like it to show zeroes for April and May, because there were no events.

It feels like there should an easy way to do this? SAS knows that it has a time axis, and it knows it's counting FailDate by month.

If there isn't an easy way I know I can calculate the counts myself, and then force in rows with count=0 and then user a series plot to plot the line. But I'd be happy if there were some option like how PROC TABULATE can use preloadfmt order=data missing=0 to force in zeroes for empty categories.

In reality I'll have a spaghetti plot, with one line for each group.

Thanks.

Re: impute zeroes when plotting frequencies with time axis

ballardw — Mon, 13 Aug 2018 19:30:19 GMT

Add some dates to the data for those months, summarize something to get counts and use a different statistic such as sum and/or the Freq= variablename

topic impute zeroes when plotting frequencies with time axis in Graphics Programming

impute zeroes when plotting frequencies with time axis

Re: impute zeroes when plotting frequencies with time axis