About skipper

skipper · ‎05-01-2012

I have code %let idn = b; %let Strike = 149; %productData; %MACRO productData; DATA product_%idn; set home.euro_c_%strike_1109; %idn_datetime = Options_Trade_date_Time; %idn_price = options_trade_price; keep %idn_datetime %idn_price; RUN; %MEND; which does not work, how do I get the %idn and %strike into my code? Also if I have sets of %idn and %strike like b, 149 c, 148 d, 147 . . m, 138 I dont want to manually change the %let statements and then re-run then code, how can I run the macro over and over again, one time for each set of %idn and %strike ?

skipper · ‎05-01-2012

I have this rather long peice of code and was wondering how it could be reduced down to take up less lines of code. data home.euro_f_1109; set home.euro_f_1109; if trade_date_time <= "22JUN11:16:00:00"dt then group=1; else if "22JUN11:17:00:00"dt <= trade_date_time <= "23JUN11:16:00:00"dt then group=2; else if "23JUN11:17:00:00"dt <= trade_date_time <= "24JUN11:16:00:00"dt then group=3; else if "24JUN11:17:00:00"dt <= trade_date_time <= "25JUN11:00:00:00"dt then group=4; else if "26JUN11:17:00:00"dt <= trade_date_time <= "27JUN11:16:00:00"dt then group=5; else if "27JUN11:17:00:00"dt <= trade_date_time <= "28JUN11:16:00:00"dt then group=6; else if "28JUN11:17:00:00"dt <= trade_date_time <= "29JUN11:16:00:00"dt then group=7; else if "29JUN11:17:00:00"dt <= trade_date_time <= "30JUN11:16:00:00"dt then group=8; else if "30JUN11:17:00:00"dt <= trade_date_time <= "01JUL11:16:00:00"dt then group=9; else if "01JUL11:17:00:00"dt <= trade_date_time <= "02JUL11:00:00:00"dt then group=10; else if "05JUL11:00:00:00"dt <= trade_date_time <= "05JUL11:16:00:00"dt then group=11; else if "05JUL11:17:00:00"dt <= trade_date_time <= "06JUL11:16:00:00"dt then group=12; else if "06JUL11:17:00:00"dt <= trade_date_time <= "07JUL11:16:00:00"dt then group=13; else if "07JUL11:17:00:00"dt <= trade_date_time <= "08JUL11:16:00:00"dt then group=14; else if "08JUL11:17:00:00"dt <= trade_date_time <= "09JUL11:00:00:00"dt then group=15; else if "10JUL11:17:00:00"dt <= trade_date_time <= "11JUL11:16:00:00"dt then group=16; else if "11JUL11:17:00:00"dt <= trade_date_time <= "12JUL11:16:00:00"dt then group=17; else if "12JUL11:17:00:00"dt <= trade_date_time <= "13JUL11:16:00:00"dt then group=18; else if "13JUL11:17:00:00"dt <= trade_date_time <= "14JUL11:16:00:00"dt then group=19; else if "14JUL11:17:00:00"dt <= trade_date_time <= "15JUL11:16:00:00"dt then group=20; else if "15JUL11:17:00:00"dt <= trade_date_time <= "16JUL11:00:00:00"dt then group=21; else if "17JUL11:17:00:00"dt <= trade_date_time <= "18JUL11:16:00:00"dt then group=22; else if "18JUL11:17:00:00"dt <= trade_date_time <= "19JUL11:16:00:00"dt then group=23; else if "19JUL11:17:00:00"dt <= trade_date_time <= "20JUL11:16:00:00"dt then group=24; else if "20JUL11:17:00:00"dt <= trade_date_time <= "21JUL11:16:00:00"dt then group=25; else if "21JUL11:17:00:00"dt <= trade_date_time <= "22JUL11:16:00:00"dt then group=26; else if "22JUL11:17:00:00"dt <= trade_date_time <= "23JUL11:00:00:00"dt then group=27; else if "24JUL11:17:00:00"dt <= trade_date_time <= "25JUL11:16:00:00"dt then group=28; else if "25JUL11:17:00:00"dt <= trade_date_time <= "26JUL11:16:00:00"dt then group=29; else if "26JUL11:17:00:00"dt <= trade_date_time <= "27JUL11:16:00:00"dt then group=30; else if "27JUL11:17:00:00"dt <= trade_date_time <= "28JUL11:16:00:00"dt then group=31; else if "28JUL11:17:00:00"dt <= trade_date_time <= "29JUL11:16:00:00"dt then group=32; else if "29JUL11:17:00:00"dt <= trade_date_time <= "30JUL11:00:00:00"dt then group=33; else if "31JUL11:17:00:00"dt <= trade_date_time <= "01AUG11:16:00:00"dt then group=34; else if "01AUG11:17:00:00"dt <= trade_date_time <= "02AUG11:16:00:00"dt then group=35; else if "02AUG11:17:00:00"dt <= trade_date_time <= "03AUG11:16:00:00"dt then group=36; else if "03AUG11:17:00:00"dt <= trade_date_time <= "04AUG11:16:00:00"dt then group=37; else if "04AUG11:17:00:00"dt <= trade_date_time <= "05AUG11:16:00:00"dt then group=38; else if "05AUG11:17:00:00"dt <= trade_date_time <= "06AUG11:00:00:00"dt then group=39; else if "07AUG11:17:00:00"dt <= trade_date_time <= "08AUG11:16:00:00"dt then group=40; else if "08AUG11:17:00:00"dt <= trade_date_time <= "09AUG11:16:00:00"dt then group=41; else if "09AUG11:17:00:00"dt <= trade_date_time <= "10AUG11:16:00:00"dt then group=42; else if "10AUG11:17:00:00"dt <= trade_date_time <= "11AUG11:16:00:00"dt then group=43; else if "11AUG11:17:00:00"dt <= trade_date_time <= "12AUG11:16:00:00"dt then group=44; else if "12AUG11:17:00:00"dt <= trade_date_time <= "13AUG11:00:00:00"dt then group=45; else if "14AUG11:17:00:00"dt <= trade_date_time <= "15AUG11:16:00:00"dt then group=46; else if "15AUG11:17:00:00"dt <= trade_date_time <= "16AUG11:16:00:00"dt then group=47; else if "16AUG11:17:00:00"dt <= trade_date_time <= "17AUG11:16:00:00"dt then group=48; else if "17AUG11:17:00:00"dt <= trade_date_time <= "18AUG11:16:00:00"dt then group=49; else if "18AUG11:17:00:00"dt <= trade_date_time <= "19AUG11:16:00:00"dt then group=50; else if "19AUG11:17:00:00"dt <= trade_date_time <= "20AUG11:00:00:00"dt then group=51; else if "21AUG11:17:00:00"dt <= trade_date_time <= "22AUG11:16:00:00"dt then group=52; else if "22AUG11:17:00:00"dt <= trade_date_time <= "23AUG11:16:00:00"dt then group=53; else if "23AUG11:17:00:00"dt <= trade_date_time <= "24AUG11:16:00:00"dt then group=54; else if "24AUG11:17:00:00"dt <= trade_date_time <= "25AUG11:16:00:00"dt then group=55; else if "25AUG11:17:00:00"dt <= trade_date_time <= "26AUG11:16:00:00"dt then group=56; else if "26AUG11:17:00:00"dt <= trade_date_time <= "27AUG11:00:00:00"dt then group=57; else if "28AUG11:17:00:00"dt <= trade_date_time <= "29AUG11:16:00:00"dt then group=58; else if "29AUG11:17:00:00"dt <= trade_date_time <= "30AUG11:16:00:00"dt then group=59; else if "30AUG11:17:00:00"dt <= trade_date_time <= "31AUG11:16:00:00"dt then group=60; else if "31AUG11:17:00:00"dt <= trade_date_time <= "01SEP11:16:00:00"dt then group=61; else if "01SEP11:17:00:00"dt <= trade_date_time <= "02SEP11:16:00:00"dt then group=62; else if "02SEP11:17:00:00"dt <= trade_date_time <= "03SEP11:00:00:00"dt then group=63; else if "06SEP11:00:00:00"dt <= trade_date_time <= "06SEP11:16:00:00"dt then group=64; else if "06SEP11:17:00:00"dt <= trade_date_time <= "07SEP11:16:00:00"dt then group=65; else if "07SEP11:17:00:00"dt <= trade_date_time <= "08SEP11:16:00:00"dt then group=66; else if "08SEP11:17:00:00"dt <= trade_date_time <= "09SEP11:16:00:00"dt then group=67; else if "09SEP11:17:00:00"dt <= trade_date_time <= "10SEP11:00:00:00"dt then group=68; else if "11SEP11:17:00:00"dt <= trade_date_time <= "12SEP11:16:00:00"dt then group=69; else if "12SEP11:17:00:00"dt <= trade_date_time <= "13SEP11:16:00:00"dt then group=70; else if "13SEP11:17:00:00"dt <= trade_date_time <= "14SEP11:16:00:00"dt then group=71; else if "14SEP11:17:00:00"dt <= trade_date_time <= "15SEP11:16:00:00"dt then group=72; else if "15SEP11:17:00:00"dt <= trade_date_time <= "16SEP11:16:00:00"dt then group=73; else if "16SEP11:17:00:00"dt <= trade_date_time <= "17SEP11:00:00:00"dt then group=74; else if "18SEP11:17:00:00"dt <= trade_date_time <= "19SEP11:16:00:00"dt then group=75; else if "19SEP11:17:00:00"dt <= trade_date_time <= "20SEP11:16:00:00"dt then group=76; else if "20SEP11:17:00:00"dt <= trade_date_time <= "21SEP11:16:00:00"dt then group=77; else if "21SEP11:17:00:00"dt <= trade_date_time <= "22SEP11:16:00:00"dt then group=78; else group=0; run;

skipper · ‎04-30-2012

so this code data data_set (drop = i); do i = 1 to 20000; X = rand('NORMAL',0,1); output; end; run; proc sql; select mean(x) into :mx from data_set; select count(x) into :cx from data_set; quit; data want (keep=T RT) / view=want; set data_set (keep=x); if _n_ > 1 then do i = 1 to _n_ - 1; set data_set (rename=(x = lagged_x)) point=i; T = _n_ - i; RT = (x - &mx) * (lagged_x - &mx) / (&cx - T); output; end; run; proc means data=want sum; OUTPUT OUT=want3; var RT; class T; run; data want4; set want3; where _STAT_="MEAN" and T > 0; keep T RT; run; gives me what I want, but it took 6 minutes to compute, and thats only for 20,000 observations, which is not large enough I need at least 2 million.

skipper · ‎04-30-2012

seems to be working, need to run some more checks

skipper · ‎04-30-2012

I dont understand, like data data_set (drop = i); do i = 1 to 200; X = rand('NORMAL',0,1); output; end; run; proc sql; select mean(x) into :mx from data_set; select count(x) into :cx from data_set; quit; data want2 (keep=T RT) / view=want2; set data_set (keep=x); if _n_ > 1 then do i = 1 to _n_ - 1; set data_set (rename=(x = lagged_x)) point=i; T = _n_ - i; RT = (x - &mx) * (lagged_x - &mx) / (&cx - T); output; end; run; proc means data=want2 sum; var RT; produces only 123 rows? it should be R(1) to R(200)?

skipper · ‎04-30-2012

looks interesting,... but data data_set (drop = i); do i = 1 to 200; X = rand('NORMAL',0,1); output; end; run; proc sql; select mean(x) into :mx from data_set; select count(x) into :cx from data_set; quit; data want (keep=T RT); set data_set (keep=x); if _n_ > 1 then do i = 1 to _n_ - 1; set data_set (rename=(x = lagged_x)) point=i; T = _n_ - i; RT = (x - &mx) * (lagged_x - &mx) / (&cx - T); output; end; run; creates a table with 19,900 observations!.. it should only have 200 observations, vector X has size 200, so it should be T = 1, 2, 3, 4, .... 200 and R(1), R(2), ... R(200).

skipper · ‎04-30-2012

thanks it seems data data_set (drop = i); do i = 1 to 200; X = rand('NORMAL',0,1); output; end; run; data data_set; set data_Set; obsN = _N_; run; proc timeseries data=data_set out=out outcorr=timedomain; id obsN interval=day; var X; run; works well, but it only gives the first 20 rows for the outcorr=timedomain whereas I need all of the rows. I need it to find R(1), R(2), ... R(200), Is there some way to specifiy how many rows the 'outcorr=<>' creats in the dataset.

skipper · ‎04-30-2012

Trying to compute the 'autocorrelations' for time series data set, 5 million observations of a single variable X. R(T) = SUM ( X - mean(X) ) * ( LagT(X) - Mean(X) ) / (N - T) must be computed and summed for each lag times T=1,2,3,4,.....,N. hence the 5million columns I dont want to create 5,000,000 columns, its my poor coding knowledge of a better way. but i need the compute the above equation R(1), R(2), .... R(5,000,000). So all i really need is just a two column data set with variables 'X' and 'R(T)' with 5 million observations. Any ideas how this could be done?

skipper · ‎04-30-2012

HI, I have code to compute 'N' means, in my code N=200. data data_set (drop = i); do i = 1 to 200; X = rand('NORMAL',0,1); output; end; run; proc sql; select mean(x) into :mx from data_set; select count(x) into :cx from data_set; quit; %macro compute; data data_set; set data_set; %do n=1 %to &cx -1 ; r&n = (X - &mx)*(lag&n(X) - &mx) ; %end; run; %mend; %compute PROC MEANS DATA=data_set mean; OUTPUT OUT=want; run; ************************************** BUT, in my real data, N=5,000,000 !! (and the above code becomes to slow with N>20,000.) How can this be done on such large 'N' ?

skipper · ‎04-30-2012

thanks that's really useful, it gives me some insight into how to code in SAS, really useful answer, thanks

skipper · ‎04-30-2012

Hi, I want to use a LAG function in a 'do' loop but the code does not work, any ideas why? data test_data; input X; datalines; 0.2 0.1 -0.3 0.1 0.1 -0.1 0.2 ; run; data autocorrelation; set test_data; do n=1 to 3 r&n = (X - 0.0429)*(lag&n(X) - 0.0429) ; run; Also, the numeric value 0.0429 used is the mean of 'X', but mean(X) does not work, so how can I code in the mean value of X and not have to hard code is numeric value.

skipper · ‎03-26-2012

thanks I will try this and report back

skipper · ‎03-26-2012

Just ran it on my large data sets A - 1.5 million obs B = 21k obs data set 'want' gets done in 0.42 seconds, WOW, amazing! and for some reason 'a_pricea now gets shown fine in the output of 'want' for these large data sets, strange?

skipper · ‎03-26-2012

seems to work on the small example data set, although variable 'a_price' seems to not show up in the 'want' table?

skipper · ‎03-26-2012

10 minutes runing and still going, will have to end this one

Online Status	Offline
Date Last Visited	‎09-01-2015 07:11 AM

how to shorten a merge code of many tables?

Macro, get variable into code

call macro function many times

run code on startup

open a file on startup

Re: macros, inserting %let strings into code

Re: macros, inserting %let strings into code

Re: macros, inserting %let strings into code

Re: macros, inserting %let strings into code

Re: macros, inserting %let strings into code

macros, inserting %let strings into code

reducing the size of my code

Re: working with large data sets

Re: working with large data sets

Re: working with large data sets

Re: working with large data sets

Re: working with large data sets

Re: working with large data sets

working with large data sets

Re: do loop with LAG function

do loop with LAG function

same trade times - large data sets

same trade times - large data sets

same trade times - large data sets

same trade times