About gergely_batho

gergely_batho · ‎03-06-2015

Yes: Standard procedure to re-fit the winning model on the entire time series and then use that model to forecast.

gergely_batho · ‎03-06-2015

Here's an SQL version. Of course, instead of (time-29) you need to use (time-29*60), or the intnx() function. data have; do id=1 to 1000; time+ranuni(123); output; end; run; proc sql; create table want as select l.id, l.time, r.id as id2, r.time as time2 from have as l, have as r where r.time between l.time-29 and l.time-31; quit;

gergely_batho · ‎03-06-2015

Hi Tao, Could you post an example of data what you have, and what results you want? What you describe is a self-join? For example, if there are 3 records within the 29-31 minute interval, you want 3 rows generated for the current row? If I understand you correctly, you can do thin with an SQL self-join, or (if you have lots of records) after sorting the data set by time, you need one carefully written datastep to accomplish this.

gergely_batho · ‎03-05-2015

Is this a homework? data want; numConsec=2;/*set this to 1, 2 or 4 ...*/ do until(eof1 and eof_1); do i=1 to numConsec; if not eof1 then do; set have(where=(a=1)) end=eof1; output; end; end; do i=1 to numConsec; if not eof_1 then do; set have(where=(a=-1)) end=eof_1; output; end; end; end; do until(eof0); set have(where=(a=0)) end=eof0; output; end; stop; run; not tested!

gergely_batho · ‎03-05-2015

Are you using: View - Workspace Layout - Side by Side ? Still the view changes, but maybe a bit easier to follow.

gergely_batho · ‎03-05-2015

I think you will need to use the filename(), fopen(), fread(), fget() functions, and check return codes. Here is an example: data urls; input url:$100.; datalines; http://support.sas.com/techsup/ http://support.sas.com/techsup/service_intro.html http://support.sas.com/techsup/ ; run; data want; set urls; length text $ 10; rc1=filename('urlref',url,'URL','lrecl=3200'); if rc1=0; fid=fopen('urlref','S'); if fid~=0; rc2=fread(fid);/*reading 1 line*/ if rc2=0; rc3=fget(fid, text, 10);/*reading 10 characters*/ rc4=fclose(fid); putlog rc1= fid= rc2= rc3= text= rc4= _ERROR_=; run;

gergely_batho · ‎03-05-2015

You want reorder your dataset so that first 1s and -1s come alternately, and at the end all the 0s ? data want; do until(eof1 and eof_1); if not eof1 then do; set have(where=(a=1)) end=eof1; output; end; if not eof_1 then do; set have(where=(a=-1)) end=eof_1; output; end; end; do until(eof0); set have(where=(a=0)) end=eof0; output; end; stop; run;

gergely_batho · ‎03-05-2015

After the input statement we need to put a variable name. Otherwise the created dataset (test) does not have columns. data test; infile rawdata obs = 10; input someColName $1000.; put _infile_; run; But I think @KurtBremser original suggestion was to use the data step generated by the PROC IMPORT: modify it by including an obs=10 to the infile statement.

gergely_batho · ‎03-04-2015

- Decreasing guessingrows (if possible) - compress=yes How much time does it take to simply copy this file with operating system command? The speed of import depends largely on the speed of the hard drive.

gergely_batho · ‎03-03-2015

List all your variables (that are not aggregated) in the group by clause. proc sql; create table work.Panel_purchase as select id, week, retailer, product_code, sum(quantity) as quantity, sum(total_price) as total_price from work.Purchase_all group by id, week, retailer, product_code, id2, month, col3, col4 /*etc...*/ order by id, week, retailer, product_code; quit;

gergely_batho · ‎03-03-2015

You can use PROC HPSPLIT if you have SAS/STAT SAS/STAT(R) 13.2 User's Guide: High-Performance Procedures

gergely_batho · ‎03-03-2015

This means, when the k-means algorithm recalculates cluster memberships, there is no change compared to the previous iteration. SAS/STAT(R) 13.2 User's Guide

gergely_batho · ‎03-03-2015

I augmented your code, so you can see in the log, what is happening. I also put an output statement at the end of the data step (this changes nothing in this case, just instead of implicit output, we use explicit output) data c; putlog '1:' _all_; set b; putlog '2:' _all_; set a; putlog '3:' _all_; set d; putlog '4:' _all_; putlog 'OUTPUT will be here'; output; run;

gergely_batho · ‎03-03-2015

Yes. At the first data step iteration: 1. SAS reads table b, b="c" - but no output to table c 2. SAS reads table a, b="a" - but no output to table c 1. SAS reads table b, b="e" - now there is an output to table c

gergely_batho · ‎03-03-2015

By the way, this is the usual way, how a data step terminates. Run this code: data class2; putlog 'BEFORE ' _all_; /*_N_=20 is printed here*/ set sashelp.class; /*exit here, when _N_=20 - no more obs*/ putlog 'AFTER' _all_; run;

Online Status	Offline
Date Last Visited	‎02-02-2017 01:26 AM

Re: LIFEREG: hazard function with SAS exponential distribution is not ...

Re: How much memory is needed? Or how to reduce memory requirements?

Re: How to model Y when a lot of 0 values come from small values taken...

Re: proc hpforest

Re: proc hpforest

Re: proc hpforest

Re: sashadoop connectivity

Re: Unexpected clusters from PROC CORR data using PROC CLUSTER

Re: Contour plot of quadratic regression model

Re: Panel Model to Disaggregate Combined Effects (Probability)

Re: How to process by process (first. last.) with in multiple groups -...

Re: How to process by process (first. last.) with in multiple groups -...

Clustering Effect Evaluation

Re: Likelihood in PROC NLMIXED with the warning of Hessian matrix not ...

Likelihood in PROC NLMIXED with the warning of Hessian matrix not posi...

Re: Clustering Effect Evaluation

Re: removing a password from a library

Re: How to model Y when a lot of 0 values come from small values taken...

Re: Panel Model to Disaggregate Combined Effects (Probability)

Re: Subset '30'-'33' (character variable)

Re: Time Series Model Hold Out Data and Rerunning the model?

Re: How can I create a table based on a readtime Column?

Re: How can I create a table based on a readtime Column?

Re: problem of sorting

Re: Screen Moving

Re: Skip Error When Reading from URL

Re: problem of sorting

Re: Effective way to import huge CSV file

Re: Effective way to import huge CSV file

Re: aggregating data by

Re: How to do the decison tree in SAS?

Re: Proc fastclus - convegence criterian

Re: Multiple set statements

Re: Multiple set statements

Re: Multiple set statements