Because SAS can stores date values as numbers (number of days after 01jan1960, or negative values for prior to 01jan1960), the answer is yes, you can interpolate dates.
I presume you want to take, for each longitude, it's relative position between longitudes with know dates, and apply that same relative progress to the time span between the known dates, as in:
proc format ;
invalue indate 'missing'=. other=[date9.];
run;
data have;
input Longitude Date :indate9. Quantity;
format date date9. ;
datalines;
24.5 24May2018 340000
25.5 missing 340000
26 missing 320000
27 missing 300000
28 missing 280000
28.5 missing 240000
29 missing 210000
30 missing 180000
30.3 31Dec2018 175000
run;
data want (drop=_:);
do _n=1 by 1 until (date^=.);
set have;
end;
_startdate=lag(date);
_ndays=dif(date); /* dif(date) means date-lag(date) */
_startlong=lag(longitude);
_difflong=dif(longitude);
do _i=1 to _n ;
set have;
if _difflong^=. and date=. then date = _startdate + _ndays*(longitude-_startlong)/_difflong;
output;
end;
run;
But please be aware that the interpolated date values above need not be integers, even though displaying them with the date9. format will look the same as if they were integers. For instance the underlying date value for 24May2018 is 21328. But if you had a value of, say 21328.99, and you applied the date9. format to its display, it would still appear as 24MAY2018.
You could make all the interpolated values into integers, if you wish. But if you do so, I'd suggest the FLOOR function (instead of round, or ceiling), just to be consistent with the date9. format.
