About slchen

slchen · ‎05-20-2015

proc sql; select * from have group by id having count(*)=1 or missing(visit); quit;

slchen · ‎05-20-2015

data have; input id $10. visit cycle $ date yymmdd10.; format date yymmdd10. ; cards; 100010001 1 CYCLE-1 2014-09-15 100010001 2 CYCLE-2 2014-10-27 100010001 3 CYCLE-3 2014-11-27 100010002 1 CYCLE-1 2014-10-13 100010002 2 CYCLE-2 2014-11-24 100010003 1 CYCLE-1 2015-02-08 100010003 1 CYCLE-1 2015-04-08 100010004 1 CYCLE-1 2015-04-28 100020001 1 CYCLE-1 2014-12-04 100020001 2 CYCLE-2 2015-01-15 100020001 3 CYCLE-3 2015-03-02 100020001 4 CYCLE-4 2015-04-13 100020002 1 CYCLE-1 2015-03-18 100020002 2 CYCLE-2 2015-04-29 100030002 1 CYCLE-1 2014-11-03 100030002 2 CYCLE-2 2014-12-22 100030002 3 CYCLE-3 2015-02-09 100030002 4 CYCLE-4 2015-03-23 100030003 1 CYCLE-1 2015-05-04 140010004 1 CYCLE-1 2015-02-11 140010004 2 CYCLE-2 2015-03-27 ; run; proc sql; select b.*,ifn(intck('week',a.date,b.date)>6,1,0) as flag from have a right join have b on a.id=b.id and b.visit-a.visit=1 order by id,visit; quit;

slchen · ‎05-19-2015

PROC SQL; CREATE TABLE work.test AS SELECT catx(' ', DISTINCT acct_ID,acct_num) AS CUST_ID FROM work.test_account ORDER BY CUST_ID; QUIT;

slchen · ‎05-19-2015

You could use merge or update statement

slchen · ‎05-17-2015

You could try this: proc sql; create table want as select a.*,b.* from dataset1 a,dataset2 b where compged(a.name1,b.name2,'L')<=1000; quit;

slchen · ‎05-15-2015

If I understand your question correctly, try this: data want(keep=company date total_dir); set have; by company date notsorted; retain term total_dir; if first.company or ceo_dummy=1 then do; term=tenure; total_dir=0; end; if tenure>term then total_dir+1; if last.company then output; run;

slchen · ‎05-15-2015

data have; input text $50.; cards; The generalized general idea Hon. member some text ; run; data want; set have; _text=prxchange('s/Hon.|some|the|general\b//i',-1,text); run;

slchen · ‎05-15-2015

proc sort data=SecondSampling; by sample; run; data want; merge FirstSampling SecondSampling; by sample; run;

slchen · ‎05-13-2015

data have; input (v1-v6)($) date mmddyy10. num; format date mmddyy10.; cards; BELTWAY REGION CDV CONNECT BUSINESS QTY 5/3/2015 442.1057 BELTWAY REGION CDV CONNECT BUSINESS QTY 5/10/2015 421.6068 ; run; data want(rename=(_num=num _date=date)); set have; do i=0 to 6; if i=0 then do; _date=date; _num=num; num=num/7; output; end; else do; _date=intnx('day',date,i); _num=num; output; end; end; format _date mmddyy10.; drop num date i; run;

slchen · ‎05-13-2015

%let measureSummed = &measureTy || &_sum; %let measureSummed=%sysevalf(%sysfunc(translate(&measureSummed,+,||))); %put &measureSummed;

slchen · ‎05-07-2015

data test; input id jan feb mar apr; array month Jan--Apr; length flag $15.; if nmiss(of month(*))=0 then flag='All'; else if nmiss(of month(*))=3 then flag='First'; else if nmiss(of month(*))=1 then do; if whichc(.,of month(*)) in (1,4) then flag='Consecutive'; else flag='Two Months'; end; else if nmiss(of month(*))=2 then flag='Two Months'; cards; 1 732 750 748 790 2 760 . . . 3 750 . 765 . 4 666 . . . 5 777 779 880 . 6 850 . 779 . ; run;

slchen · ‎05-06-2015

proc sql; create table want as select * from test where id not in (select distinct id from test where gestage not between 20 and 43) ; quit;

slchen · ‎05-06-2015

&&code&j first resolve to &code1, then &code1 resolve to 146.

slchen · ‎05-05-2015

It depends on how you deal with missing value. If there is missing value, you either don't calculate or omit it by this way, total_day=sum(last_day,-start_day,-hold_resume,hold_start);

slchen · ‎05-05-2015

Try this: proc sql; create table p as select *, avg(proportion) as grp_avg from(select agencyid, count(case answer when 'yes' then 1 else 0 end) as yes_count, ((count(case answer when 'yes' then 1 else 0 end))/24) as numerator, count(distinct unitid) as denominator, ((calculated numerator/calculated denominator)*100) as proportion from test1 where questionid in ('Q1','Q2','Q3') group by agencyid); quit;

Online Status	Offline
Date Last Visited	‎07-12-2017 03:27 PM

Re: How to exclude some names when use proc sql to create macrolist?

Re: Counting character variable in array with substr

Re: DATETIME variable stored as a string

Re: Moving observations in the required position

Re: Number Format to Date Format

Re: Extracting keywords and corresponding number

Re: Extract conditional information from text

Re: %do without using %to

Re: How do I combine multiple columns with missing data and prioritizi...

Re: Counting the number of occurrences for a grouped variable

Re: Last 5 characters from string(irrespective of the length)

Re: Replacing " " with a "_" for variable names in a given file

Re: ODS graphics with more than 5000 points have been suppressed

Re: How to exclude some names when use proc sql to create macrolist?

Re: Moving observations in the required position

Re: SAS QUERY

Re: sas query

Re: how to trim to column and concatenate in PROC SQL?

Re: Merge / Append two differnt columns into one column

Re: Proc SQL & Cost Function in one step

Re: counting the number of observations conditional on another observa...

Re: deleting different elements from a text

Re: Special Binding Data

Re: How do I create 6 more rows using one row using SAS

Re: Let Compute a Value rather than Store a String

Re: Summary

Re: How to remove a group of observations based on conditions

Re: Macro Resolution

Re: Date Calculation

Re: proc sql performing repeat calculations and creating a table of t...