BookmarkSubscribeRSS Feed
deleted_user
Not applicable
Dear colleges,

Let me describe the situation, please.

There's an extract from the code where I've got a macro MonteCarlo which runs MC simulations for a specified k, and a macro Back_testing which calls the macro MonteCarlo in a cycle for 250 times.

With the last part of MonteCarlo being commented (proc univariate, data-step and proc append) the code runs successfully. But when I'm trying to run it as whole the process never stops (Task Status Box says that proc sql / proc iml is running), although the messages in the log indicate that all the iterations were completed successfully (no errors, no warnings), even %put statement (the last one in the code) was executed. The code has no command to be back to the macro MonteCarlo. Why does it go? And why does it stop here?..

I tried to relpace the procedures commented (proc univariate, data-step and proc append) with the simple data step like
data VALUE_AT_RISK; VaR = 1; run;
but it also failed (without this data step the code is executed, but placing it in the code makes execution process never stop).

And the last note: when I run the whole code in smaller cycles (for k = 2 to 100, then for k = 101 to 200 and lastly for k = 201 to 251) everything goes right as well.

I would greatly appreciate it if you could comment on the case.

Thank you in advance,
Elena


Here's the extract:


%macro MonteCarlo(k);
%let _date = %scan(&_dates251,&k,' ');
%let _fut_date = %scan(&_dates251,&k-1,' ');
proc sql noprint;
select count(Currency_Cd) into :_count from POSITION where Day_Dt = &_date;
select '"' || trim(Currency_Cd) || '"' into :_curs separated by ' ' from POSITION where Day_Dt = &_date;
select '"' || trim(Currency_Cd) || '"' into :_curs1 separated by ' ' from POSITION where Day_Dt = &_date and Currency_Cd ne 'USD';
select Currency_Rate format=best12. into :_USD_rate from RATES where Day_Dt = &_date and Currency_Cd = 'USD';
select Currency_Rate format=best12. into :_USD_rate_actual from RATES where Day_Dt = &_fut_date and Currency_Cd = 'USD';
select Position format=best12. into :_USD_position from POSITION where Day_Dt = &_date and Currency_Cd = 'USD';
select sum(Position) into :_port_value from POSITION where Day_Dt = &_date;
run;
%let _unquoted_curs = %sysfunc(compress(&_curs,'"'));
%let _unquoted_curs1 = %sysfunc(compress(&_curs1,'"'));
proc iml;
use LN_RATES1 (firstobs=&k obs=%eval(&k+249) keep=&_unquoted_curs1);
read all into R;
C = corr(R);
H = root(C);
avg = R[:,];
dif = R - repeat(avg,250,1);
ssq = dif[##,];
std = sqrt(ssq/249);
std = T(std);
use POSITION1 (firstobs=&k obs=&k keep=&_unquoted_curs1);
read all into pos;
pos = T(pos);
use RATES_USD_BASED1 (firstobs=&k obs=&k keep=&_unquoted_curs1);
read all into rates;
rates = T(rates);
posNom = pos/(rates#&_USD_rate);

portLoss = J(&_num_scenarios,1,1);
do j=1 to &_num_scenarios;
y = normal(repeat(0,&_count-1,1));
z = T(H)*y;
futRates = rates#exp(std#z#sqrt(10));
futUSDRate = (&_h_USD_edge - &_l_USD_edge) * ranuni(0) + &_l_USD_edge;
futPortValue = futUSDRate#(sum(posNom#futRates)+&_USD_position/&_USD_rate);
portLoss = futPortValue - &_port_value;
end;

create SCENARIOS from portLoss [colname = 'Portfolio_Loss'];
append from portLoss;

use RATES_USD_BASED1 (firstobs=%eval(&k-1) obs=%eval(&k-1) keep=&_unquoted_curs1);
read all into futRatesAct;
futRatesAct = T(futRatesAct);
futPortValueAct = &_USD_rate_actual#(sum(posNom#futRatesAct)+&_USD_position/&_USD_rate);
portLossAct = futPortValueAct - &_port_value;

create ACTUAL_PORT_LOSS from portLossAct [colname = 'Actual_Portfolio_Loss'];
append from portLossAct;
abort;
run;

/*proc univariate data = SCENARIOS noprint;
output out = VALUE_AT_RISK p1 = VaR;
run;

data bt;
merge VALUE_AT_RISK ACTUAL_PORT_LOSS;
run;

proc append base = BCKTST_RES data = bt; run;*/
%mend MonteCarlo;

%macro Back_testing;
%do k = 2 %to 251;
%put &k;
%MonteCarlo(&k);
%end;
%mend Back_testing;

data BCKTST_RES;
VaR = .;
Actual_Portfolio_Loss = .;
run;

%Back_testing;

%put 'THE END';
3 REPLIES 3
deleted_user
Not applicable
At a very quick glance I would say that your problem is that you don't have a QUIT; statement terminating your PROC SQL. This is something I have fallen foul of on several occasions and a quick scan of your code couldn't find a quit.

I can't comment on the PROC IML but this could be something similar, or the PROC is running inside the PROC SQL.

Hope this helps.
deleted_user
Not applicable
Thank you a lot for the reply, Deadhead.

Indeed, I've got the same general idea of the case: some of the steps within the macro don't exit properly. This might be the reason why execution process somehow goes back to them.

So I put QUIT statement into PROC SQL, and also closed all the data sets within PROC IML explicitly. The code runs successfully when only PROC SQL and PROC IML are included in the macro. However, it still fails to finish executing when I try to add, for instance, PROC UNIVARIATE. I put QUIT statement into PROC UNIVARIATE also. But it doesn't help.

Could somebody give any suggestions of the ways to terminate those procedures (SQL, IML, UNIVARIATE, APPEND) and data-step explicitly? I tried some, but it's still in question, unfortunately...
Any other ideas would be also appreciated. Thank you for the attention 🙂
deleted_user
Not applicable
I think you should try to find a way to debug the macro part maybe, rather than stop the program. I would try using on mprint and mlogic options.

sas-innovate-2024.png

Today is the last day to save with the early bird rate! Register today for just $695 - $100 off the standard rate.

 

Plus, pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

Click image to register for webinarClick image to register for webinar

Classroom Training Available!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 3 replies
  • 826 views
  • 0 likes
  • 1 in conversation