About tmm

Tom · ‎02-20-2015

Use the technique posted by @data_null_; here: Just adjust to use the ~ modifier instead of the : modifier to force the quotes. Also did you want TAB delimited like your example or COMMA delimited (which is what CSV means). data _null_; set SASHELP.class; file log dlm=',' dsd; if _n_ eq 1 then link names; put (_all_)(~); return; names: length _name_ $32; do while(1); call vnext(_name_); if upcase(_name_) eq '_NAME_' then leave; put _name_ ~ @; end; put; return; run;

tmm · ‎03-14-2014

That is the ticket. Thanks.

tmm · ‎03-11-2014

Thanks. These are both great options.

PGStats · ‎01-29-2014

Only a single query needed for this : DATA HAVE; infile datalines missover; input CLM_AUD_NBR $ (proc1-proc20) ($); array p proc:; do procNo = 1 to dim(p); proc = p[procNo]; if not missing(proc) then output; end; keep clm_aud_nbr procNo proc; datalines; 111 11900 99212 . 11111 . . . . . . . . . . . . 22222 . . 222 82270 99214 . . 11900 . . . . . . . . . . . . 33333 . 333 17000 17003 99213 . . . 11111 . . . . . . . . . . . 44444 444 96372 99213 . . . . . . 22222 . . . . . . . . . . 55555 555 92012 92083 . . . . . 22222 . . . . . . . . . . . 666 17000 17003 99213 . . . . . . 33333 . . . . . . . . . 777 90836 99215 . . . . . . . . 44444 . . . . . . . . 888 92004 92081 . . . . . . . . . 55555 . . . . . . . 999 90836 99215 . . . . . . . . . . 66666 . . . . . . 121 92014 92083 . . . . . . . . . . . 77777 . . . . . 131 36410 99214 . . . . . . . . . . . . 88888 . . . . 141 92012 92083 . . . . . . . . . . . . . 99999 . . . 151 92012 92083 . . . . . . . . . . . . . . . . . 161 92014 92083 . . . . . . . . . . . . . . . . . 171 G0008 Q2036 . . ; DATA HAVE2; input procFirst $ procScnd $; datalines; G0008 99999 G0008 Q2036 G0008 92083 G0008 33333 17000 17003 17000 99213 90836 99215 96372 99213 ; proc sql; create table deniedClaims as select distinct hf.clm_aud_nbr from have as hf inner join have as hs on hf.clm_aud_nbr=hs.clm_aud_nbr and hf.procNo<hs.procNo inner join have2 as h2 on hf.proc=h2.procFirst and hs.proc=h2.procScnd; select * from deniedClaims; quit; PG

tmm · ‎01-29-2014

Let me put together a data have. Might make better sense.

CTorres · ‎01-27-2014

Hi Patrick, You are right. The instructuion I meant was a MERGE instead a SET. I already Edited the original message. My solution gives only four observations and asumes that both datasets are sorted by proc1 proc2. My test program follows: DATA TABLE1; input PROC1 PROC2 $ EDIT; Datalines; 100 G0427 0 102 0311T 0 102 36010 0 102 76998 0 ; run; DATA TABLE2 ; input CLAIM PROC1 PROC2 $ NET1 NET2 ADJD1 ADJD2 ADJDT1 $10. ADJDT2 $10.; Datalines; 111 100 3032 5 5 111 873 10/10/2013 10/10/2013 222 100 G0427 5 5 111 873 10/11/2013 10/11/2013 333 102 0 19 15 111 873 10/12/2013 10/12/2013 444 102 36010 20 20 111 873 10/13/2013 10/13/2013 ; run; Data need; merge table2 (in=t2) table1(drop=edit in=t1); by proc1 proc2; flag=t1 and t2; if t2; run; Regards,

art297 · ‎01-21-2014

; You might find it a little easier to do using the macro a group of us presented at SAS Global Forum last year. You can download the macro at: A Better Way to Flip (Transpose) a SAS Dataset - sasCommunity The following shows your example code and the macro call that will produce the desired result: libname cci "c:\art"; DATA cci.phyclms2; input CLM_AUD_NBR PROC_CD SBMT_CHRG_AMT ALLW_AMT NET_PD_AMT; cards; 111 92083 150 65.41 65.41 111 99213 105 92.12 67.12 111 92014 275 166.95 141.95 222 99213 155 96.3 71.3 222 99214 175 130.94 105.94 333 99213 98 62.71 37.71 333 99204 250 104.95 79.95 ; %transpose(data=cci.phyclms2, out=cci.phyclms3, by=CLM_AUD_NBR, var=PROC_CD SBMT_CHRG_AMT ALLW_AMT NET_PD_AMT, sort=yes, newid=row)

tmm · ‎07-23-2013

Thanks Vincent. I deleted it, though not worried about anyone finding it.

Haikuo · ‎07-22-2013

OK, I will just shut up and consider myself lucky. :smileysilly:

art297 · ‎07-16-2013

Then, why not use Tom's suggested code, but first do the desired calculations? e.g.: proc summary data=have; var mbr; output out=sums (drop=_:) sum=mbrsum; by strata; run; data have (drop=i);; merge have sums; array stuff(*) caths--mbr; by strata; do i=1 to dim(stuff); if not missing(stuff(i)) then stuff(i)=stuff(i)/mbrsum; end; run; Then do the flipflop.

Vince28_Statcan · ‎07-17-2013

If you use Esjackso1's approach, in order to "transpose" to a dataset like taxid perf_1 perf_4 perf_8 ... stemi_1 stemi_4... you cannot (at least not as far as I know) do so in a single transpose. A way around that is to transpose your 56 variables back in a single tall vertical dataset like taxid plofsrvcid oldvar value 1 1 perf 250 1 4 perf 110 etc. and then simply concatenate oldvar with plofsrvcid as the name of your new var and transpose back. It takes 3 step or possibly 2 if you use multiple ids in the second transpose (plotsofsrvcid and oldvar but I am not sure if it trims the strings properly). Vincent

DBailey · ‎06-24-2013

lots of columns/records to update.... Is cardiac.pos sorted or indexed? If not, then you definitely need to do that.

ballardw · ‎06-19-2013

I think you are looking for CHAR(40) to create a 40 character variable. without a width specified the default for char is 8

tish · ‎05-10-2013

I don't think you need a case statement. I'd read the character field fst_srvc_dt in as a SAS date, convert it to a Julian date, extract the day segment and subtract 1: put(substr(put(input(fst_srvc_dt, anydtdte10.), julian5.), 3), 3.) - 1 as date Note that the highest value will be 364 for 2011 and that you will get just one date variable. If you want 365 of them, I'd use PROC TRANSPOSE to transpose the data.

tmm · ‎05-02-2013

Sorry ballardw, that is just a typo. As for the error, it is posted in my message. We have had all sorts of issues with this new SAS server.

Online Status	Offline
Date Last Visited	‎09-01-2015 07:11 AM

Place quotes around column header and data

Re: Point = error

Point = error

Re: Insert into confusion

Re: Insert into confusion

Insert into confusion

Re: create a proc sql without the need for multiple proc sqls

create a proc sql without the need for multiple proc sqls

Re: how to combine 2 inner joins to proc sql

how to combine 2 inner joins to proc sql

format max date

concatenating number columns

Re: Place quotes around column header and data

Re: Point = error

Re: Insert into confusion

Re: create a proc sql without the need for multiple proc sqls

Re: how to combine 2 inner joins to proc sql

Re: Wrong data step to flag data

Re: proc transpose not outputting data needed

Re: loop usage

Re: Ridiculous Coding Process

Re: proc transpose

Re: Count and Group By issue

Re: Update Query Error

Re: Adding a column to a SAS Table

Re: Date Conversion

Re: proc format percent problem