About ArtC

ArtC · ‎02-07-2012

Remember if you use SORT with NODUPLICATES (or SORT followed by FIRST. and LAST.) that you must have a sufficient key to bring the duplicate records next to each other or they will not be eliminated. This paper has been around a while but is still important today: http://www2.sas.com/proceedings/sugi25/25/po/25p221.pdf

ArtC · ‎02-07-2012

SQL processes the entire table in memory. As the size of your tables increase you may experience performance degradation. MERGE processes a row at a time so rarely has memory limitations.

ArtC · ‎02-07-2012

A refinement on Chuck's code. proc summary data=sashelp.shoes nway; class region subsidiary; output out=counts ; run; proc print data=counts; run; Use the CLASS (no sorting required), no var statement (summary counts by default), the NWAY gives counts of the interaction. Remove the NWAY to get the counts of higher summary levels.

ArtC · ‎02-07-2012

I have not used the Teradata ACCESS, but with other instances there is no warning. I would not expect one here.

ArtC · ‎02-07-2012

I am not sure about your question, but I did notice that in your code above you have: %if %lowcase(%substr(&date,1,4))= year %then %let year_date=1; *Don't know if should put something here?; %else %let year_date=0; You may not place a comment or other statement between the %IF and its %ELSE.

ArtC · ‎01-04-2012

Good point Ksharp. If all you wanted to do was rename the variable, DATASETS would definately be a better way. As you also show, the main point is the syntax for naming a variable with non-standard characters. I have been busy writing my next book (on advanced techniques) which will be out in the spring.

ArtC · ‎01-04-2012

Just to followup on Art's reply. If the character is in the data it should not be a problem - if it is in the data and it is giving you a problem then the following will not help - it is something else again. Here is a silly example with ++ representing a Danish character which is not on my keyboard. My suggestion is to rename as soon as possible. Notice that the variable name with the special characters is quoted and followed immediately with a n. option validvarname=any; data withdanish; 'ab++c'n = 5; run; data coffeeonly; set withdanish(rename=('ab++c'n=abc)); run;

ArtC · ‎01-04-2012

You are almost certainly applying the format in the FREQ step. In FREQ the missing values are eliminated first, then the format is applied. Using the format in an earlier step (Linlin uses it in the DATA step) causes this timing to change.

ArtC · ‎08-26-2011

There are a number of ways of calculating age http://www.wuss.org/proceedings10/coders/2943_4_COD-Carpenter.pdf , however for this problem a simplier solution is available. Try the INTNX function with the 'SAME' option as the fourth argument to advance the date of birth 6 years. data students; input StudentID $ BirthDate :mmddyy10.; dob6 = intnx('year',birthdate,6,'s'); if '01jul2011'd le dob6 le '01sep2012'd then note='In Range'; format birthdate dob6 date9.; datalines; 10007 05/24/2007 73001 06/10/1997 73002 09/11/2002 10236 09/20/2004 10538 07/22/2004 10538 07/22/2005 10538 07/22/2006 10538 07/22/2007 run; proc print data=students; run;

ArtC · ‎08-22-2011

Much clearer. I am still not quite sure what we are doing with the first column, but the following makes sense to me as a next step closer. I added a title and some spanning text in a REPORT step. I also changed the ODS file extension to RTF. data types; input curr_resiStatus $ count bad Total_Bad Index participants :percent.; format participants percent.; datalines; Boarding 34 0.323529 11 -36 12% Buyer 17 0.058824 1 391 6% Livingw 4 0 0 . 1% Other 5 0.2 1 22 2% Owner 70 0.142857 10 84 25% Renting 150 0.28 42 -22 53% Supplied 1 1 1 -100 0% run; title1 'This is a title'; ODS RTF BODY='tables.rtf' style=Journal; proc report data=types nowd; columns ('Spanning Text' curr_resiStatus count bad Total_Bad Index participants) ; rbreak after / summarize; run; ODS RTF CLOSE;

ArtC · ‎08-21-2011

Without knowing what you want your table to look like it is very hard to guess what to show to you. Here is some code that changes variable names, variable labels, titles, and text. Since there is no data to summarize, TABULATE seems a less likely choice. data types; input Type $ B C D E F :percent.; label b = 'count'; datalines; Boarding 34 0.323529 11 -36 12% Buyer 17 0.058824 1 391 6% Livingw 4 0 0 . 1% Other 5 0.2 1 22 2% Owner 70 0.142857 10 84 25% Renting 150 0.28 42 -22 53% Supplied 1 1 1 -100 0% run; title1 'My Special Data'; proc print data=types; run; proc report data=types nowd; columns Type B C D E F; define c / display 'R O I'; run;

ArtC · ‎08-16-2011

You might have better luck posting this to the Statistical Procedures forum. Or check with the Stat procedure developers.

ArtC · ‎08-02-2011

Unfortunately not Eric Sun and I looked into a number of protection issues not only for the macro variables, but also for the macros themselves. We were able to come up with a few work arounds, but protecting the variables themselves is still not possible. See the paper at: http://www.lexjansen.com/pharmasug/2011/ad/pharmasug-2011-ad17.pdf

ArtC · ‎08-02-2011

Without asking why; you want a repeated character up to 100 times in the first column, but is that what you want? If the error is a listing destination error and you are not using the listing destination, turn it off ( ods listing close;) My memory says that in the listing destination a given column cannot be wider than the width of the paper (LINESIZE= system option). So increasing the LINESIZE might change the error. In the COMPUTE BAR block you multiply per_osha11 by 50, any value of per_osha11>2 (it can be up to 24) would give a total gt 100. any value of barsize>99 will give a bar with ge 100 characters.

ArtC · ‎08-01-2011

Good catch Mike. If these are indeed $4 cause of death codes then the OP must be sure to understand the usage of the : comparison modifier.

Online Status	Offline
Date Last Visited	‎06-20-2021 07:11 PM