About Keith

Keith · ‎02-01-2013

If it's for display purposes and you want to keep the id column as numeric, then just use the Zw.d format which pads the values with leading zeros. e.g. format id z5.; By the way, in your required output, you've added 4 zeros not 3.

Keith · ‎01-29-2013

Nice answer, I was working on something very similar (using PROC SUMMARY with COMPLETETYPES). The bit I was struggling with to code relatively simply was missing imputation and coping with both lag and lead values (the lag part is easy). The only problem I can see with your solution is if there is only a value for 2003 in the example data (i.e. 2001 and 2002 are missing). Using firstobs=2 will still return a missing value here. Ideally you would want to read sequential records until a value is encountered, then return that. The nature of the data means that it is not possible to go to the next BY group before a non-missing value occurs, so that isn't a worry.

Keith · ‎01-28-2013

or proc sql can do this simply... proc sql; create table want as select *, sum(Bill) as Bill_sum from have group by id; quit;

Keith · ‎01-23-2013

Here's a simple piece of code that creates a dataset of all Sundays in a year. The 'week' interval in the INTNX function returns Sunday by default. %let rpt_yr=2012; data sundays_&rpt_yr.; format sunday_date date9.; do i=0 to 52; sunday_date=intnx('week',"01Jan&rpt_yr."d,i); if year(sunday_date)=&rpt_yr. then output; end; drop i; run;

Keith · ‎01-16-2013

just run a data step with the put statement to create a new variable with the formatted values length age_band $10; age_band =put(age,$age.);

Keith · ‎01-10-2013

Here's one way using 2 arrays. I've hard coded the new variable list, but you could extract the values from the source table and store them in a macro list. data have; input Customer Var1 $ Var2 $ Var3 $ Var4 $; cards; 1 Value_A Value_B Value_C Value_D 2 Value_E Value_F Value_G Value_H 4 Value_I Value_J Value_K Value_L ; run; data want; set have; array oldvals{*} Var: ; array newvals{*} Value_A Value_B Value_C Value_D Value_E Value_F Value_G Value_H Value_I Value_J Value_K Value_L; do _i=1 to dim(newvals); do _j=1 to dim(oldvals) until(oldvals{_j}=vname(newvals{_i})); newvals{_i}=(oldvals{_j} eq vname(newvals{_i})); end; end; drop _: var: ; run;

Keith · ‎12-14-2012

PROC SUMMARY is good for this type of problem. Here's an example that returns the record with the lowest age in SASHELP.CLASS. Note that if there is more than 1 record with the same minimum value then this procedure will only return the first one (i.e. it always returns only 1 record). If you need to return all records with the minimum value then I would suggest PROC SQL, although this method is likely to be slower than PROC SUMMARY (provided there's enough memory for SUMMARY to work!) proc summary data=sashelp.class nway; output out=want1 (drop=_:) minid(age(_all_))=; run; proc sql; create table want2 as select * from sashelp.class having age=min(age); quit;

Keith · ‎12-12-2012

I would be tempted to use the INTNX function against Calendar_Month to return the first of the month, then you can just specify the exact dates in the IN clause. WHERE INTNX('MONTH',Calendar_Month,0) IN ('01Oct2011'd, '01Dec2011'd);

Keith · ‎11-30-2012

PROC STDIZE can do this for you, which saves hard coding the minimum / maximum values. (uses data set created by ) proc stdize data=have out=want method=range mult=100; var prob; run;

Keith · ‎11-23-2012

makes a good point. If you just store the raw numeric data, you can create and apply a format with the values in your post. If you are summarising the data using any of the usual SAS procs, they will automatically group by the formatted values.

Keith · ‎11-19-2012

This is clearly a homework type question. Why not try researching it first, running some code and then post a question on here if the code doesn't work?

Keith · ‎11-16-2012

If you're not sure of the variable type then use the MISSING function, although if you're assigning a new value then you obviously need to know which type to use! It's not covered in the scope of you question, but your code can be simplified using either of the 2 methods below. This saves checking for the missing value each time. if missing(AEACN) then do; if _AEACN = 1 then AEACN = .K; else if _AEACN = 2 then AEACN = .V; else if _AEACN = 3 then AEACN = .D; else if _AEACN = 4 then AEACN = .A; end; drop _AEACN; if missing(AEACN) then do; select (_AEACN); when (1) AEACN = .K; when (2) AEACN = .V; when (3) AEACN = .D; when (4) AEACN = .A; otherwise end; end; drop _AEACN;

Keith · ‎11-16-2012

very simple, = ' ' refers to a character missing value. You're applying it to a numeric variable, so it is converting the character missing value to a numeric missing value. You can either change ' ' to . which is the default missing numeric value, or use missing(varname) which applies to both character and numeric variables (and will also pick up special missing numeric values such as .K)

Keith · ‎11-09-2012

These solutions assume the data has been transposed, although I think a table structure that has Player_Id and Player_Number as the variables is preferable (in which case your solutions would be good).

Keith · ‎11-09-2012

Given the structure of your data, my instinct is to create a format from your list of player_id's and then read across your main dataset to see if the formatted value exists. Here's an example using 3 player_id's.. data have1; input player_id1-player_id3; cards; 555 600 740 ; run; data have2; input player_id; cards; 555 650 700 740 ; run; data id_fmt; set have2 (rename=(player_id=start)); retain fmtname 'id_fmt' type 'N'; label='check'; run; proc format cntlin=id_fmt; run; data want; set have1; array pl{*} player_id1 - player_id3; array plchk{*} player_check_id1 - player_check_id3; do i=1 to dim(pl); plchk{i}=put(pl{i},id_fmt.)='check'; end; drop i; run;

Online Status	Offline
Date Last Visited	‎09-11-2015 12:18 PM

Re: Regarding If and then statement

Re: Merging 3 datasets

Re: Regading if and then

Re: Retaining Values in a Do Loop

Re: Using column wildcard (:) with retain statement in a datastep

Re: Select observations that meet a criteria

Re: Select observations that meet a criteria

Re: How to read Input data

Re: Output next,or next and next,or next and next and next,hahaha

Re: Re-Arrange Variables

Re: Find the missing values in the year

Re: Increment row based on change in record

Re: Dealing with 32 character restrictions

Re: calculating averages on floating period

Need help in matching the obs from dataset to excel

Re: How to change the name and format of PROC IMPORT variables

Re: Adding zeros for the number...

Re: Find the missing values in the year

Re: Total Sum vertically

Re: need help in date fuction

Re: Age varible as character variable and with the formatted values. I...

Re: Creating variables from variable values

Re: get the minimum value of one variable

Re: SAS limitation with IN operator

Re: How to convert probability to 0-100 scale

Re: SAS Proc and sorting options

Re: Query for dummy coding with females as a reference group.

Re: Character values have been converted to numeric values when enteri...

Re: Character values have been converted to numeric values when enteri...

Re: I need help checking if each observation contains a record from a ...

Re: I need help checking if each observation contains a record from a ...