About Tom

Tom · ‎10-01-2011

For something so simple you can also just use a DATA _NULL_ step and not bother with macro language. %let type=color; %let varlist=; data _null_; if "&type" = 'color' then call symput('varlist','red blue'); if "&type" = 'pattern' then call symput('varlist','solid stripe'); run; data data1 (keep= &varlist); set datalib.data1; run;

littlestone · ‎10-02-2011

I apologize for repling late. Thank you all for great help.

LeBoulet · ‎11-02-2011

Yes, this option tells SAS where to find rbrowser and it works. But it wasn't the issue. What I have not found is how to tell rbrowser where to find Firefox: it always tries to find it on `hostname`, which in my example was pg02 instead of 192.168.31.112, or even more conveniently 127.0.0.1 (no need to tweek the configuration for each computer with this one).

mrug12 · ‎09-30-2011

Thanks Tom , It is really help full.

Ksharp · ‎09-30-2011

You do not use Hash Table correctly. See the followig code of example. BTW. I am not sure this method will be more efficient than proc sort; data temp; input name age; cards; 23 45 67 89 12 56 ; run; data _null_; *load dataset temp into Hash Table; if 0 then set temp; declare hash ha1(dataset:'temp',ordered: 'a'); ha1.definekey('name','age'); ha1.definedata('name','age'); ha1.definedone(); *output Hash Table to a dataset; ha1.output(dataset:'want'); run; *now see the data has already been sort by name age; proc print data=want;run; Ksharp

Pritish · ‎09-27-2011

@Art and @Tom - Thanks for sharing the knowledge. @Peter, Well it depends on the business rules and requirements. I was writing some code and came up with this and was not able to think clearly on the solution. So that's the reason I came up with this question.

art297 · ‎09-27-2011

Tom, Many of us use procs import and export because we figure that SAS would have made the task easier and with more capabilties than we could have accomplished on our own. Unfortunately, that isn't always the case as we have discovered with procs import and export. Hey, we could also accomplish building a neural net, logistic regression, or what have you in a datastep as well but, I think, most of us would simply use the procs included with Enterprise Miner and SAS/STAT. Me thinks that your comment should be taken as a wake up call to those who are responsible for procs import and export. Some of our colleagues and I are doing a presentation at SGF2012 related to proc import, specifically about importing tables that one might copy from the web. We, too, had to build code in a datastep because proc import didn't provide the necessary capabilities. Obviously, a much better job could have been done if we had access to the actual source code.

Tom · ‎09-26-2011

As Art says add PATIENT to your TABLES statement. If you want it as a list then use the LIST option. Otherwise you will get a grid with #rows equal to the number of patients and #cols equal to the number of categories. TABLES patient*diagnosis / list ;

stan · ‎02-19-2013

Dear nishant nair, your reply is interesting for me , but how to generalize it to all tables in SAS outputs from different PROCs ? Thanks.

jeeth79usa · ‎04-07-2012

Hello All, I need help with the similar situation. Thanks to you all for your postings. It was very informative. I was wondering on how to handle the proc http, if my URL is a variable in a data set called HTTP. I have used the url access method, but I was getting an error saying invalid server response. I need to run all the url links one by one and get the values stored on the server (some name), and send those values to the name variable back to my work data set. I was thinking to export the file to the temp folder on unix, and then use the filename in temp; filename out temp; Please advise. I have used your method here: My environment is UNIX: proc sql noprint; select count(*) into :httpcnt from work.URL_Test where substr(http,1,5)='http:'; quit; %put &httpcnt; data _null_; set Test; call symput("cpe",trim(variable)); call symput("date_time",trim(variable)); run; %let url = "http://host:80abcdetcetc%nrstr(&cpe)%nrstr(%&)incidentDate=%nrstr(&date_time)"; filename http_get "/test/test/http.csv"; %macro get_http; /*%do i=1 %to &httpcnt;*/ proc http out=http_get url=&url method="get" ; run; /*%end;*/ %mend; %get_http; Btw, This method worked for me when I took just a single link from the variable to test its out. It worked great, and created forllowing valuesin the out=http_get.csv file. Please let me know if its possible to get only "name I am looking for"? are there any options to remove the tags?? I need to run all the links, and include the results to my data set (in name variable). Thanks a lot in advance. *****http_get.csv file. <html> <head> <title>adcdef</title> </head> <body> <h1>name I am looking for</h1> </body> </html>

Ksharp · ‎09-21-2011

Hi. Peter Long time no see. It is good to see you again. As Peter said, in SAS there is already a informat to read these data. data want; input a $ ; b=input(substr(a,2),julian5.); c=put(b,yymmdd10.); put a= b= c=; cards; 110272 110232 ; run; a=110272 b=18534 c=2010-09-29 a=110232 b=18494 c=2010-08-20 Ksharp

Cynthia_sas · ‎10-21-2011

Hi, As Tim pointed out, WIDTH= option of PROC REPORT only has an impact on LISTING output. Since the OP said he wanted HTML output, the WIDTH= option would not work. Instead, a combination of changing the default split character (a /) and using CELLWIDTH should get him what he wants. In the attached screen shots, examp_width1.jpg shows code that used LISTING and HTML. Without WIDTH=, the label 'Gender' does behave as you describe, but only for the LISTING output. As you can see, in the HTML output, the label appears correctly. However, because the split character is defined as / (the default), the label string for NAME splits at the slash in the label. With the SPLIT= redefined, and an explicit CELLWIDTH, the header for NAME is as desired and the header for SEX is fine without a WIDTH= option. The units for CELLWIDTH can be specified as pixels, points, inches or centimeters. I usually don't mess with pixels and I just go with inches as the unit. In addition to WIDTH=, there are other PROC REPORT options that only work in the LISTING destination, these are: FLOW, PANELS, BOX, DOL, DUL, OL, UL, SKIP, HEADLINE, HEADSKIP, WIDTH, SPACE, LS, PS. See this Tech Support note for more information: http://support.sas.com/kb/2/549.html http://support.sas.com/kb/23/671.html http://support.sas.com/rnd/base/ods/templateFAQ/report1.html cynthia

Ksharp · ‎09-25-2011

Thanks . Tom. Educated it.

bncoxuk · ‎09-20-2011

Hi art, your code is also correct. As Tom said, I should have put capital letters in the libname and memname statement. Thank you very much.

Mr_Breeze · ‎09-19-2011

Thanks, Tom!! It does work across multiple pages, and parses everything nicely. Once I transpose the account file and join it to the company file, it will be lined up perfectly for my purposes. I greatly appreciate all of the help!

Online Status	Online
Date Last Visited	yesterday