About TomKari

TomKari · ‎08-06-2019

When you look at your server list with "View > Servers", do both servers appear under the "Servers"? If so, both servers are available in the same metadata server environment. If not, you'll need to switch between different metadata profiles. Tom

TomKari · ‎08-02-2019

I'm glad this helped you. Just to confirm, yes, the code I supplied would only be used on the variables that are defined as 4 byte integer in Teradata. Any variables that are 8 byte integer or any length of floating point would not have a problem (although with floating point, you need to understand at what point you'll start to lose low-end digits with integers). Tom

TomKari · ‎08-01-2019

Don't worry about formats; they don't change the underlying value of the variable. Let's check the easiest thing first; run this code on your data. If there's a value that's not within limits, the error diagnostic will show you the line and the variable. Tom data _null_; set abc; if col1 > 2100000000 | col1 < -2100000000 then abort; if col2 > 2100000000 | col2 < -2100000000 then abort; if col3 > 2100000000 | col3 < -2100000000 then abort; if col4 > 2100000000 | col4 < -2100000000 then abort; if col5 > 2100000000 | col5 < -2100000000 then abort; if col6 > 2100000000 | col6 < -2100000000 then abort; if col7 > 2100000000 | col7 < -2100000000 then abort; if col8 > 2100000000 | col8 < -2100000000 then abort; if col9 > 2100000000 | col9 < -2100000000 then abort; if col10 > 2100000000 | col10 < -2100000000 then abort; if col11 > 2100000000 | col11 < -2100000000 then abort; if col12 > 2100000000 | col12 < -2100000000 then abort; if col13 > 2100000000 | col13 < -2100000000 then abort; if col14 > 2100000000 | col14 < -2100000000 then abort; if col15 > 2100000000 | col15 < -2100000000 then abort; if col16 > 2100000000 | col16 < -2100000000 then abort; if col17 > 2100000000 | col17 < -2100000000 then abort; if col18 > 2100000000 | col18 < -2100000000 then abort; if col19 > 2100000000 | col19 < -2100000000 then abort; if col20 > 2100000000 | col20 < -2100000000 then abort; if col21 > 2100000000 | col21 < -2100000000 then abort; if col22 > 2100000000 | col22 < -2100000000 then abort; if col23 > 2100000000 | col23 < -2100000000 then abort; if col24 > 2100000000 | col24 < -2100000000 then abort; if col25 > 2100000000 | col25 < -2100000000 then abort; if col26 > 2100000000 | col26 < -2100000000 then abort; if col27 > 2100000000 | col27 < -2100000000 then abort; if col28 > 2100000000 | col28 < -2100000000 then abort; if col29 > 2100000000 | col29 < -2100000000 then abort; if col30 > 2100000000 | col30 < -2100000000 then abort; if col31 > 2100000000 | col31 < -2100000000 then abort; if col32 > 2100000000 | col32 < -2100000000 then abort; if col33 > 2100000000 | col33 < -2100000000 then abort; if col34 > 2100000000 | col34 < -2100000000 then abort; if col35 > 2100000000 | col35 < -2100000000 then abort; if col36 > 2100000000 | col36 < -2100000000 then abort; if col37 > 2100000000 | col37 < -2100000000 then abort; if col38 > 2100000000 | col38 < -2100000000 then abort; if col39 > 2100000000 | col39 < -2100000000 then abort; if col40 > 2100000000 | col40 < -2100000000 then abort; if col41 > 2100000000 | col41 < -2100000000 then abort; if col42 > 2100000000 | col42 < -2100000000 then abort; if col43 > 2100000000 | col43 < -2100000000 then abort; if col44 > 2100000000 | col44 < -2100000000 then abort; if col45 > 2100000000 | col45 < -2100000000 then abort; if col46 > 2100000000 | col46 < -2100000000 then abort; if col47 > 2100000000 | col47 < -2100000000 then abort; if col48 > 2100000000 | col48 < -2100000000 then abort; if col49 > 2100000000 | col49 < -2100000000 then abort; if col50 > 2100000000 | col50 < -2100000000 then abort; if col51 > 2100000000 | col51 < -2100000000 then abort; if col52 > 2100000000 | col52 < -2100000000 then abort; if col53 > 2100000000 | col53 < -2100000000 then abort; if col54 > 2100000000 | col54 < -2100000000 then abort; if col55 > 2100000000 | col55 < -2100000000 then abort; if col56 > 2100000000 | col56 < -2100000000 then abort; if col57 > 2100000000 | col57 < -2100000000 then abort; if col58 > 2100000000 | col58 < -2100000000 then abort; if col59 > 2100000000 | col59 < -2100000000 then abort; if col60 > 2100000000 | col60 < -2100000000 then abort; if col61 > 2100000000 | col61 < -2100000000 then abort; if col62 > 2100000000 | col62 < -2100000000 then abort; if col63 > 2100000000 | col63 < -2100000000 then abort; run;

TomKari · ‎08-01-2019

Teradata probably uses a 4 byte integer format, and that is the largest number that can be contained within that format. Whereas SAS uses an 8 byte floating point format for numbers, and can store numbers with up to 15 digits. If you need to store the number accurately in Teradata, you'll need to change the variable assignment in Teradata to either 8 byte integer (if that exists in Teradata), or eight byte floating point. This will be a job for your database administrator, so start by talking to them. Tom

TomKari · ‎07-24-2019

This is the general format that SAS uses to pass back an error message thrown by Sybase IQ, so you need to figure out what's going wrong in the Sybase IQ environment. In addition to the other comments, which are spot-on, I notice that you appear to be using the ODBC interface to Sybase IQ. SAS has released a native connection engine for IQ, which in my experience worked much better. If possible, I suggest that you acquire that and use it. You may see performance and reliability benefits, as IQ works significantly differently from regular ODBC-type databases. One problem that I recall was that SAS by default was asking for way too few records to be returned at a time, which was hammering performance. Note that all of these comments apply to when I was last working with SAS and IQ, circa 2012, so things may have changed. Tom

TomKari · ‎07-16-2019

MR variables are tricky, although they seem to be straightforward. I suggest you set up two datasets; one of respondents with the appropriate variables, and the other of responses, with the appropriate variables. Tabulations on these two datasets should get you what you want. Tom

TomKari · ‎07-09-2019

If the problem is that there are more than 50 occurrences, a prior step to count the max number of occurrences should fix you up. Tom data _null_; file "C:\Some file.txt"; Rec = "987 123 456 789 012 345 678 901 1.55042962215361e-05 456 789 012 345 678 901 345"; put Rec; Rec = "654 123 456 789 012 345 678 901 123 456 789 012 345 1.55042962215361e5 901 345"; put Rec; Rec = "321 123 456 789 -1.55042962215361e-05 345 678 901 123 456 789 012 345 678 901 345"; put Rec; Rec = "098 123 456 789 012 345 678 901 123 456 -1.55042962215361e05 012 345 678 901 345"; put Rec; run; data _null_; retain _MaxVarCount 0; infile "C:\Some file.txt" end=_LastRec; input; _VarCount = countw(_infile_, " "); if _VarCount > _MaxVarCount then _MaxVarCount = _VarCount; if _LastRec then call symput("VarCount", strip(put(_VarCount, best15.))); run; data MyData(drop=_:); infile "C:\Some file.txt"; length _InRec $32767; length var01-var&VarCount. 8; array _RecVars{&VarCount.} var01-var&VarCount.; input; _InRec = _infile_; _VarCount = countw(_InRec, " "); do _i = 1 to _VarCount; _RecVars{_i} = input(scan(_InRec, _i, " "), best32.); end; run;

TomKari · ‎07-05-2019

Yes, that's exactly what I was asking about! I think this should do what you need; just put the code that you want executed conditionally between the %if and the %end statements (I used a proc print step). Tom data _NULL_; rc = 0; if day(date()) = day('2jun2019'd) then rc = 1; CALL SYMPUT('DateFlag', PUT(rc,1.)); stop; run; %if &DateFlag = 1 %then %do; /* Code to be run conditionally */ proc print data=sashelp.class; run; %end;

TomKari · ‎07-04-2019

I just did some experimenting, and I can't see any way to get the conditional logic into the exported SAS code. However, in the program that I am running conditionally, after I've run it, there is a tab called "Conditional Logic Log". It shows how EG transitions between your specified condition and a flag that EG uses to decide whether to run the program. If you're comfortable with macro logic, you should be able to use that to derive the code that you need. If not, post it here and someone will be able to give you an example. Tom

TomKari · ‎07-04-2019

In addition to all of the above suggestions, which make great sense, if you have full control over your SAS environment you can relocate your WORK library similar to how you moved your UTILLOC. Tom

TomKari · ‎07-04-2019

I've modified it to input the number using the BEST format. See if this advances your progress. Tom data _null_; file "C:\Some file.txt"; Rec = "abc 123 456 789 012 345 678 901 1.55042962215361e-05 456 789 012 345 678 901 345"; put Rec; Rec = "def 123 456 789 012 345 678 901 123 456 789 012 345 1.55042962215361e5 901 345"; put Rec; Rec = "ghi 123 456 789 -1.55042962215361e-05 345 678 901 123 456 789 012 345 678 901 345"; put Rec; Rec = "jkl 123 456 789 012 345 678 901 123 456 -1.55042962215361e05 012 345 678 901 345"; put Rec; run; data MyData(drop=_:); infile "C:\Some file.txt"; length _InRec $32767; length var01-var50 8; array _RecVars{50} var01-var50; input; _InRec = _infile_; _VarCount = countw(_InRec, " "); do _i = 1 to _VarCount; _RecVars{_i} = input(scan(_InRec, _i, " "), best32.); end; run;

TomKari · ‎07-02-2019

Here's a code snippet that should help you with the issue of getting the variable number of columns into variables. Note that the first data step is just to create some text data to test on. Tom data _null_; file "C:\some file.txt"; Rec = "abc 123 456 789 012 345 678 901 123 456 789 012 345 678 901 345"; put Rec; Rec = "def 123 456 789 012 345 678 901 123 456 789 012 345 678 901 345"; put Rec; Rec = "ghi 123 456 789 012 345 678 901 123 456 789 012 345 678 901 345"; put Rec; Rec = "jkl 123 456 789 012 345 678 901 123 456 789 012 345 678 901 345"; put Rec; run; data MyData(drop=_:); infile C:\some file.txt'; length _InRec $32767; length var01-var50 $ 16.; array _RecVars{50} var01-var50; input; _InRec = _infile_; _VarCount = countw(_InRec, " "); do _i = 1 to _VarCount; _RecVars{_i} = scan(_InRec, _i, " "); end; run;

TomKari · ‎06-25-2019

Here's a quick code snippet that might help you figure out the problem: FILENAME CCREP "some file.txt"; data _null_; infile CCREP; input; file log; put _infile_; if _n_ = 5 then stop; run; First of all, you should get some diagnostics on the characteristics of the dataset. Second, it will print the contents of the first five records, so you can get some idea of if the lengths make sense and if the contents are correct. From there, you (and we, if necessary) can try to figure out what's up. Tom

TomKari · ‎06-19-2019

It's well worth saving them to use in the future. You can use the "Save as" button. Make sure you give them a meaningful name, and you can also organize them by creating and using directories. Tom

TomKari · ‎06-18-2019

Personally, my head always starts to hurt when I get into too many macro functions and variables. It's less sophisticated, but I find a data step clearer. Tom %let aux = 201803; data _null_; OriginalDate = mdy(input(substr("&aux.", 5, 2), best2.), 1, input(substr("&aux.", 1, 4), best4.)); ModifiedDate = intnx("month", OriginalDate, -11, "same"); call symput("aux1", put(ModifiedDate, yymmn6.)); run; %put &=aux.; %put &=aux1.;

Online Status	Offline
Date Last Visited	‎07-11-2025 09:14 PM