@PeterClemmensen nice. Please note that you'll need start and end of string anchors to avoid matching 'AF10' or 'F999' for example. @CynthiaWei I'm not sure if it is intended but by looping like this you always overwrite your flag with the examination of D30, you may like to reformulate more precisely what you intend to do. Shall all values D01-D30 match the requirement or just any of them? ... View more

Hi, I'd probably determine the values in macro like this: %macro loop(); %local value i; %let value="F01"; %do i=2 %to 99; %if not (20<=&i. and &i.<=29) %then %let value=&value. "F%sysfunc(putn(&i,z2.))"; %end; &value. %mend loop; %let range=%loop(); %put &range; then in your datastep: if D(i) in (&range.) then Flag=1; ... View more

Hi, you need to stop the loop after a replacement is made and stop overwriting new_policy_number over and over again until you try to replace with 'MANUALCEDED' which always fails.   data test_set ; infile datalines truncover; input policy_number $100. ; datalines; Q2 2017 Manual Entry 000123400 Manual Manual_fix15 Manual_fix22Q1 Manual All Data 0098765401CededManual 0098765401Manual 0011223301 CededManual ; run; data chk; set test_set; length new_policy_number $100; retain new_policy_number ; array nn[4] $50 _temporary_ ('MANUALENTRY' 'MANUAL' 'CEDEDMANUAL' 'MANUALCEDED'); do i=1 to dim(nn); new_policy_number = upcase(compress(tranwrd(policy_number,'_',''))); if index(new_policy_number,nn[i]) then do; new_policy_number = tranwrd(new_policy_number,nn[i], '') ; leave; end; end; drop i; run; Please also note that in obs #7 you're first replacing with manual since it comes first in your array and not 'cededmanual'     ... View more

N1ce that should do the trick 😊 ... View more

Hi,   none I know of, but a workaround could be to estimate the dataset size before a memory load? ... View more

Hi,   it might be possible to retrieve the files using the pipe command which would be much faster. (untested) %let path=; filename files pipe "ls -haltr *.* ""%unquote(&path.)"" " lrecl=32767; data test; infile files truncover; input files $char1000.; put files=; run; ... View more

of course, it's by -1 My mistake ... View more

Hi, loop from last to first and compare manually: do i=dim(temp1) to 1; if vvalue(temp1[i]) eq coalescec(of temp1[*]) then lastnonmiss1=i; end; ... View more

Hi, This isnt' a procedure but a SAS macro that runs the R analyses in the R program to redirect the results to SAS. Also the data is transferred with proc import/export between the environments. The paper and the macro are 10 years old (!) but the procedures used in SAS are still available with no significant update (if any) so I'm not sure what kind of update you expect? ... View more

One way is to use the dictionary tables to select the variables you want   demo: %let names=; PROC SQL;/*libname and memname for consultation purpose only*/ SELECT libname, memname, name INTO :lib,:mem,:names SEPARATED BY ' ,' FROM sashelp.vcolumn WHERE libname EQ upper('sashelp') AND memname EQ upper('geoexm') AND name LIKE upper('name%') ; QUIT; %symdel lib mem; %put &=names; PROC SQL outobs=10/*limit results to 10 observations*/; SELECT &names. FROM sashelp.geoexm ; QUIT; Adapt the code to your requirements. ... View more

Hi, check your "not equal" operator. In SAS it is written "NE" or "^=" ... View more

Hi,   not directly, since several formats could contain the value 'asdf'. In a dataset, the function vformat returns the format that is associated with the specified variable. But I bet you already know this. You could write out the formats and search for the ones that contain the value you look for.   ods listing; proc format cntlout=work.myformats; run; %let seek=%upcase(asdf); %let fmtnames=; PROC SQL noprint; SELECT distinct FMTNAME INTO :fmtnames SEPARATED BY ' ' FROM myformats WHERE index(upcase(strip(label)),"&seek.") OR index(upcase(start),"&seek.") OR index(upcase(end),"&seek.") ORDER BY fmtname ; QUIT; %put Formats containing "&seek.": &=fmtnames.;     ... View more

Hi,   I'd try first to import the *.txt file into a dataset to see if I could use the data structure to create the output I want If this is not feasible I think I'd import it as bulk in a first step, use regular expressions or text analysis function (index, findw) in a second step to extract the data I need and finally create the output in a third step. Since this is mainly data triggered you need to know what the data exactly looks like and if the data structure may change in future (maxim 3)   ... View more