About Tom

Kurt_Bremser

What is actually created at compile time is the temporary .sas7bdat.lck file. And this file is renamed to .sas7bdat when the step terminates even in case of ERROR if no previous dataset exists.

Kurt_Bremser

For further clarification: RESOLVE can only resolve macro elements, but it can not execute the resulting code. If macro elements within a macro are meant to be created/changed by non-macro code contained in the macro, this will not work as intended, as the non-macro code will never run. CALL EXECUTE, OTOH, feeds code into the execution queue. But you need to take care when a macro is involved, as macro triggers will be resolved immediately while all other code can only run after the current step finishes. Therefore, %NRSTR is needed to mask macro code during the insertion into the execution queue; it will then be resolved when code is fetched from the queue.

yabwon

Use the MacroArray package. 1) Documentation 2) Article %macro inf_value(ttt,information_value,numeric_Vars,Ind_Failure); %put [&sysmacroname.] &numeric_Vars.; data _null_; &numeric_Vars.=42; put _all_; run; %mend; %loadPackage(macroArray) /* macroarray */ %array(varsList[*] a_Shir_SmlHov_Pigur a_Shir_AKM_HZRHIUV a_Shir_Vetek_age_Balut a_CDO_ExCC_mx_pos_prob ,vnames=Y ,macarray=Y) /* loop over macroarray calling %inf_value() macro */ %do_over(varsList,phrase=%nrstr( %inf_value(ttt,information_value,%varsList(&_i_.),Ind_Failure); )) /* clean up after using macroarray */ %deleteMacArray(varsList,macarray=Y) Bart

andreas_lds

SAS Studio, Enterprise Guide, Display Manager and even the sas-addon for VS Code display library contents always in alphabetical order. But if you should have access to your work-folder with Windows Explorer, too, so that you use it to display datasets in the required order.

Tom

I am not sure what you mean by "stack". But it looks like you want to resolve a series of macro variables with the same base name and a numeric suffix. Why not use SYMGET() function? data mvtemp； length var $32 ; var = symget(cats('var',_n_)); input note $8.; cards; X Xx YYY ; Or since it looks like the values of the VARxx macro variables have digit strings perhaps you want to use SYMGETN() instead? data mvtemp； value = symgetn(cats('var',_n_)); input note $8.; cards; X Xx YYY ;

Tom

As long as the length of the macro variable is less than 32K bytes you can use normal data step. data want; do i=1 to countw(symget('list'),'+'); length var_name $32 ; var_name = scan(symget('list'),i,'+'); output; end; run; If the string is (or could be) longer then you might want to use a macro %DO loop instead. Which will require defining a macro. %macro list2ds(dsname,list); %local i ; data &dsname; length var_name $32 ; %do i=1 %to %sysfunc(countw(&list,+)); var_name = "%scan(&list,&i,+)"; output; %end; run; %mend list2ds; %list2ds(dsname=want,list=&list)

Kally80

Hi Tom, Thanks for your response. I too realized the issue quite late. Infact what I have done is that I introduced a length statement [like how you mentioned] and got the code working fine. Thanks again for taking the time to get back. Regards, Kailly

DLehwidi

Wow, this seems to be working, thank you so much, Could you please try searching from multiple columns like on my example. where you search on multiple variables

js5

We could probably get this set up, but it would require considerably more adjustments than just changing which ODBC driver is being used by our ODBC libname. Right now, we have the SQL server set up as an external data catalogue in DBX, and we can just copy the data to corresponding databricks schemas using CTAS statements. The goal of using ODBC to load to DBX directly would be to potentially eliminate the need for the SQL Server altogether. Having to go the S3 route would need to be analysed in terms of how much resources it would cost to set up.

ballardw

Don't forget the Picture formats also supports the MULTIPLIER option to do the multiplication (or division by 100 if using a value of 0.01 for multiplier) though it always takes me some trial and error to get the right multiplier for the use I need as I don't use it often enough.

Kurt_Bremser

If bp_group is to be a Boolean value (yes/no), you can shorten it to bp_group = (120 <= Systolic_BP0 < 130);

Tom

Which comes first LEAVE or ARRIVE? Are we talking about a trip where you leave before you arrive? Or hospital stays where you leave after you arrive? What is the logic for deciding which variable to change? What is the logic for deciding when having a different year is a problem. What if they arrived on Dec 31 and left on Jan 1? What if the duration was actually more than 365 days? What if changing the year makes an invalid date? Feb 29 only exists in leap years.

RobPratt

@Ksharp Sorry, I had oversimplfiied the code, which I have now corrected. The corrected line is: set VARS {IDS} init VARS_ALL;

DocMartin

I've spoken with my colleague about his absurd data formatting. He said he'll try to fix it. So, for now, this task is moot. Thanks to all who took the time to respond.

FreelanceReinh

Fully agree with Tom. I would expect that those values would be judged unequal even within Teradata, i.e., the transfer to SAS has not caused the difference. Note that the same sort of numeric precision issues can easily occur in SAS, too. Tom's solution would be the remedy here as well. Example: 288 data _null_; 289 total_amount=7+0.11+0.27; 290 mean_amount =7+0.27+0.11; 291 if total_amount > mean_amount then flag=1; 292 else flag=0; 293 rflag=(round(total_amount,1e-9) > round(mean_amount,1e-9)); 294 put (_all_) (=best32.); 295 put (_all_) (hex16. /); 296 run; total_amount=7.38 mean_amount=7.38 flag=1 rflag=0 401D851EB851EB86 401D851EB851EB85 3FF0000000000000 0000000000000000 Note that even the BEST32. format conceals the tiny difference (of 2**-50=8.881...E-16) between total_amount and mean_amount, which arose from just adding three numbers in different orders and which is reliably revealed by the HEX16. format.

Online Status	Offline
Date Last Visited	Saturday