PROC FREQ is the SAS component designed to count things. It looks like this problem can be solved in one step, using built-in features. data have ; input CODE $                    SPECIALTY $      ; cards ;    PS                                OBGYN PS                                 SURG PS1                               OBGYN PS1                                 ANES PS1                                  NEPH PS2                                OBGYN PS2                                  NEPH ; proc freq data=have noprint ; by code ; format code $2. ; tables specialty / out=want( rename = (count=OCCURRENCES) drop = percent ) ; run ; raveena wrote: Hi , I need a help to get an deisred output as below, (Need to consider PS1 ,PS2  as PS) CODE                      SPECIALTY PS                                OBGYN PS                                 SURG PS1                               OBGYN PS1                                 ANES PS1                                  NEPH PS2                                OBGYN PS2                                  NEPH OUTPUT SHOULD BE APPEAR AS BELOW, CODE                       SPECIALTY           OCCURRENCES PS                               OBGYN                            3 PS                                 SURG                            1 PS                                  ANES                            1 PS                                   NEPH                           2 Thanks in Advance.. ... View more

Good that you learned. The person who prepared the file also has things to learn. The design is very bad. Were there two variable-length fields instead of one, it would have been difficult if not impossible to parse. littlestone wrote: Thank you all. I really learned a lot. ... View more

Yes, it does depend on interpretation. I can think of only (1) the INPUT statement and (2) the FREAD function. Others (PROC IMPORT, External File Interface, Enterprise Guide) arguably don't count because they simply generate code for #1. ... View more

It can be done in one statement. Edit the little pipe-separated string if necessary to get the correct offsets. data _null_ ; do cal_date = today() to today()+6 ;    batch_date =     intnx(  'day'           , cal_date           , input(  scan(  '2|1|1|1|1|3|2'                          , weekday(cal_date)                          , '|' )                   , f. ) ) ;    put (cal_date batch_date)(= downame.) ;    end ; run ; Result: cal_date=Sunday batch_date=Tuesday cal_date=Monday batch_date=Tuesday cal_date=Tuesday batch_date=Wednesday cal_date=Wednesday batch_date=Thursday cal_date=Thursday batch_date=Friday cal_date=Friday batch_date=Monday cal_date=Saturday batch_date=Monday mick_g wrote: I need to calculate a process date from an add date. for each record I have a add_date to reflect the day it  was added I need to calculate a batch date which is 1 day after the add_date except on Weekends. Tuesday batches add_dates for Sunday and Monday my code below works except for the monday batch data test_dates; set trials; cal_date = datepart (add_date); batch_date=intnx('day',cal_date,1); format batch_date yymmddp10.; run; How do I get Sunday add_dates to have batch_dates = tuesday? ... View more

Here's a way using the DATA step: data _null_ ; set sashelp.shoes ; length __name $32 ; call vnext(__name); r1c1 = vvalue(__name) ; put r1c1= ; stop ; run ; jerry898969 wrote: I need to get the value of the first variable first row of a data set without knowing the variable name.  The reason for this is I have to create a macro that will loop through many datasets and look at the first variable first row's value.  Each data set has a different name for the first variable. Any help would be appreciated. Thank you ... View more

Some people will interpret as "getting started" with SAS software. Doc@Duke wrote: They are on the main page.  However, unless you have a hi-res screen, they are off the bottom. That's a challenging trade-off.  "Getting started" is in the way of frequent users, but may be needed for newer ones. Also, for the ForumAdmin:  You may need a way to "pin" generic getting started info near the top.  Otherwise, the people who need it most won't be able to see it.  (Assuming "getting started" means to get started using the Forums rather than the growing pains we see now. Doc Muhlbaier Duke ... View more

I think you've answered your own question. $CUSTOMER is the host OS (operating system) folder or directory name. I would not call it "physical" because it's pretty abstract vis a vis the hardware implementation of the OS filespace. Only if your OS disallows dollar signs in file names is there a mystery. Here's a Windows XP example. options noxwait xsync ; x cd c:\temp ; x mkdir $CUSTOMER ; libname name '$CUSTOMER' ; data name.test ; run ; libname name clear ; Fernando wrote: Hi, I tried to find the meaning of $ sign inside the following quotes. libname name '$CUSTOMER' In my SAS class, I was taught physical location is specified inside the quotes. I explored SAS documentation but could not get a clear understanding. Could anyone of you help me? Thanks Mirisage ... View more

SAS doesn't have the DATEDIFF function. See INTCK instead. Use DTDAY as the interval if your values are datetimes rather than dates. PROC SQL and the DATA step are both suitable, but you can't mix and match. CASE is part of SQL and can't be used in the DATA step. mick_g wrote: I have the code running successful in SQL.  but I am having trouble converting into SAS. SQL code cd_1 = case when datediff (d,a.add_date, b.cd_date) ,<= 1 then 1 else 0 end cd_2 = case when datediff (d,a.add_date, b.cd_date) ,<= 2 then 1 else 0 end sas code create table test as select a.add_date           ,b.cd_date           ,cd_1 = case when datediff (d, a.add_date, b.cd_date) <= 1 then 1 else 0 end           ,cd_2 = case when datediff (d, a.add_date, b.cd_date) <= 2 then 1 else 0 end from data_333 a left join test_data b on a.client = b.client and a.id = b.id; or is it a data step; data test; set data_333 test_data; cd_1 = case when datediff (d, a.add_date, b.cd_date) , <= 1 then 1 else 0 end cd_2 = case when datediff (d, a.add_date, b.cd_date) <= 2 then 1 else 0 end ... View more

We don't seem to have a solid explanation, so I guess it's time to put this in front of Tech Support. The point is that _CHARACTER_ is *not* reserved, at least not in the way _ALL_ is. Also notice that enforcement is more or less limited to the DATA step. Here is a PROC step which is happy to permit use of variable names which are not allowed (or at least discouraged) in the DATA step. [pre] proc summary data=sashelp.class(obs=3) nway ; class sex ; output out=badnames(rename = (_freq_ = _n_ _type_=_error_) ) min(age) = _numeric_ max(age) = _character_ mode(age)= _all_ ; run ; [/pre] Of course referencing these names will be a problem. Another slightly interesting bit of evidence. When this code [pre] proc print ; var _numeric_ _character_ _char_ _all_ _notmetoo_ ; run ; [/pre] is placed in the enhanced editor, the first four specs in the VAR statement are recognized and thus colored blue rather than black. ... View more

The COALESCE function makes things somewhat more concise. [pre] proc sql; create table want as select have1.*, coalesce (h2.weight, h2_s.weight, 1) as weight from have1 as h1 left join have2 as h2 on h1.postal=h2.postal left join have2 as h2_s on substr(h1.postal,1,3)=h2_s.postal order by id; [/pre] Notice that I got rid of the alias for HAVE1 as well as the COMPRESS calls. I saw no need for them. Also, better test data: [pre] data have1; input id postal $; cards; 1 K0A3K0 2 K0A8PN 3 K0A3N0 4 K0A8T6 5 XXXXXX ; data have2; input postal $ weight; cards; K0A 0.987 K0A3K0 0.69 K0A3N0 0.74 ; [/pre] > thats fantastic, thank you so much. > > i'm not too familiar with creating tables in sql, how > would I change the code so that for anyone mnissing a > 'weight' (did not match to the lookup table) I would > set their weight value to a 1? > > Thanks. ... View more

Eliminate the WHERE clause, but move the condition into the ON clause (connected with AND). That's probably what's desired. Outer unions are guided by their ON clauses. Any accompanying WHERE clause operates as a simple filter and does not generate any "outer" rows. With inner unions, it doesn't matter if some of the conditions are coded in a WHERE clause. With outer unions it does. Here's an example. [pre] data table_L ; input id_L $ date_L ; cards ; x 1 y 2 z 3 ; data table_R ; input id_R $ date_R amount ; cards ; x 1 1.1 y 2 2.2 ; PROC SQL ; TITLE 'All Conditions in ON Clause' ; SELECT * FROM table_L LEFT JOIN table_R ON id_L EQ id_R AND date_L EQ date_R ; TITLE 'Some Conditions in WHERE Clause' ; SELECT * FROM table_L LEFT JOIN table_R ON id_L EQ id_R WHERE date_L EQ date_R ; TITLE ; QUIT ; [/pre] The first query returns [pre] id_L date_L id_R date_R amount ------------------------------------------------ x 1 x 1 1.1 y 2 y 2 2.2 z 3 . . [/pre] but the second returns [pre] id_L date_L id_R date_R amount ------------------------------------------------ x 1 x 1 1.1 y 2 y 2 2.2 [/pre] > Yes. > I think that is because you add a "where" clause. > So it will filter some obs. > You can test it without this "where" clause, to see > whether has 13098 unique . > > But the best way to find answer is to post some dummy > data and code you used,then can check it further. > > Ksharp ... View more

The JULDATE function gets you close. Use MOD to discard the year portion. [pre] data _null_ ; have = '01jul2010'd ; want = mod(juldate(have),1000) ; put want= ; run ; [/pre] > Hi! > > I am very new to SAS and am using EG. I need a > formula, or code, that will generate the day of the > year (example; for 7/1/2010, the value 181 would be > generated). Any idea how to do this? > > Much thanks! ... View more