|
On 3 fév, 21:17, Tom Abernathy <tom.aberna...@gmail.com> wrote:
> To make your macro work easier have it generate just that statements
> you want rather than the whole data step. Then embed the call to it
> in the data step.
> In your original example the name of the variable with the time (T)
> did not vary. Why is it varying in your macro version?
>
> %macro a(n);
> %local j;
> %do i=0 %to &n;
> if T-&i > 0 then d&i= mwspread/((1+_mlibor)**(T-&i)); else d&i=0;
> %end;
> %mend;
>
> data b;
> set a;
> by id_bb year month;
> %a(125);
> run;
>
> Next exercise is it to eliminate the macro completely and just code it
> using arrays.
>
> On Feb 3, 3:07 pm, ChrisG <chris.godlew...@gmail.com> wrote:
>
> > Hello all
>
> > After spending a long time searching for a clue around here without
> > success I am submitting you my (probably easy & dumb) problem
>
> > Here is what works well with my data:
>
> > data b; set a; by id_bb year month;
> > if t > 0 then d0=mmwspread/((1+libor)**(t)); else d0=0;
> > if t - 1 > 0 then d1=mmwspread/((1+libor)**(t-1));else d1=0;
> > if t - 2 > 0 then d2=mmwspread/((1+libor)**(t-2));else d2=0;
> > if t - 3 > 0 then d3=mmwspread/((1+libor)**(t-3));else d3=0;
> > if t - 4 > 0 then d4=mmwspread/((1+libor)**(t-4));else d4=0;
> > if t - 5 > 0 then d5=mmwspread/((1+libor)**(t-5));else d5=0;
> > if t - 6 > 0 then d6=mmwspread/((1+libor)**(t-6));else d6=0;
> > run;
>
> > now the issue is that in the initial data i had t = 1 to 7 (t is for a
> > time period)
>
> > now in the big dataset i have t = 1 to 126 !!!
>
> > so instead of typing 125 lines of sas code i tried to do a macro
>
> > something like this:
>
> > %macro a;
> > %do t=1 %to 126;
> > %do i=0 %to 125;
> > data b; set a; by id_bb year month;
> > if &t-&i > 0 then d&i= mwspread/((1+_mlibor)**(&t-&i)); else d&i=0;
> > output;
> > %end;
> > output;
> > %end;
> > %mend;
> > %a;
>
> > the problem is that this macro is not "appending" the d&i variable but
> > puts the newest value of d&i in the palce of the old one
> > at the end i get just the d125 in the data
> > but i want to have the d0 until d125
>
> > could you help me out please ?
>
> > thanks in advance
> > Cheers
> > CG
Thank you Tom !
You made me discover a very efficient way to do this thing
it was so fast !
and i did an error for making varying t...
it is fixed !
thank again very much !
best
CG
| 
