Any solution for this?   We got the same in VA 7.5. When looking into the css we only have a start and end color.   <![CDATA[ #ve56 .TwoColorRamp { fill-gradient-start-color: #F8F9FA; fill-gradient-end-color: #000000; } ]]> ... View more

hi Patrick, we found the same support link and checked with our SAP support team but they could not fix the issue. You think it's SAP related or we're missing a setup within the SAS configuration. As we are able to connect to SAP i'm thinking it's more SAP related? ... View more

PRESERVE_TAB_NAMES=YES would do the trick ... View more

Hi @Filipvdr    I had a look at Jim Moon's paper, and I notice that it relies heavily on the between-operator. It is a good approch to interval-joining small tables, but it is extremely slow when working with large tables, especially DBMS-tables.   The macro I mentioned in my previous post (attached) is developed to overcome that and work fast with big tables. We use it in many production jobs, where interval joins with between used to run up to 2 hours, and now run in as many minutes instead.   Given your test input, this is the call:   %fletrensinterval( tableA, tableB, tableOUT, unita=FKEY, datefirsta=FROM, datelasta=TO, unitb=FKEY, datefirstb=fromb, datelastb=tob, join=left, byvarA=ID); TableA and TableB are the input tables, and tableOUT the resulting output tables.   unita, datefirsta and datelasta are the variables holding the key to join on and the from- and to-dates i TableA. unitb, datefirstb and datelastb are the variables holding the key to join on and the from- and to-dates i TableB.   Join=left preserves all keys/intervals in TableA, even if the key is not present in TableB, or (part of) an interval in TableA  is outside corresponding intervals in TableB    ByvarA=ID is the name of a grouping variable in TableA (in this case ID), so separate joins are made for each ID value i A, if the join key (in this case FKEY) exists with more than one ID in TableA   Result:       ... View more

Hi @Filipvdr  Run the SASTRACE statement then rerun poorly performing code. Then post the log to this thread. If you can get the same performance characteristics using PROC SQL you can set the NOEXEC option. This will generate the SQL that SAS is passing to SQL Server without running the actual query..    It will help to know the relative sizes of the tables.    If I had to guess, I would say that there is an issue with indexing and statistics on the database tables. If you are using SAS functions (that cannot pass to the database) in your WHERE predicates this could cause issues, too. But, that wouldn't explain the significant performance issue you are seeing.    I have added a link to my initial response (I want to keep the suggestions in a single post). Have a look at it. You may want to submit a tech support track. They may have seen this problem in the past. ... View more

@Filipvdr  You face all sorts of timing issues when combining SAS Macro code with RSUBMIT which require quoting to resolve. I guess you know that already given the current solution you found.   The problem is that the DIS conditional Start/End does exactly that: It wraps a macro around everything in its middle. What you need to do is to write your RSUBMIT code "macro sensitive" - so different code bits need to execute depending whether the RSUBMIT is wrapped into a SAS macro or not.   You could use a conditional start/end generated macro to determine if your code is "macro wrapped" or not. Below sample code uses DIS generated code and works inside and outside the cond start transformations. /*%macro etls_conditionW2W2H81N;*/ /* %local etls_conditionTrue;*/ /* %let etls_conditionTrue = %eval(1=1);*/ /* %if (&etls_conditionTrue=0) %then*/ /* %do;*/ /* %put ETLS_DIAG: Condition flow did NOT execute, condition was 1=1;*/ /* %goto exitetls_conditionW2W2H81N;*/ /* %end;*/ /* %else*/ /* %do;*/ /* %put ETLS_DIAG: Condition flow did execute, condition was 1=1;*/ /* %end;*/ /*---- Start of User Written Code ----*/ options autosignon=yes sascmd= "!SASCMD"; rsubmit test; data _null_; call symputx('rtest','ABC'); stop; run; data _null_; if symexist("etls_conditionTrue") then call execute('%nrstr(%sysrput ltest=&rtest;)'); else call execute('%sysrput ltest=&rtest;'); stop; run; endrsubmit; %put &=ltest; /*---- End of User Written Code ----*/ /*%exitetls_conditionW2W2H81N:*/ /*%mend etls_conditionW2W2H81N;*/ /**/ /*%etls_conditionW2W2H81N;*/  If you uncomment the code currently commented then you get the DIS generated code with a conditional start/end. ... View more

Just  detail, but the char() function is better suited to this test than the index function if index(WKT12,'.') = 7   ... View more

Hi Erik, thanks for the tips..   In the end I found this solution:   Output a proc contents to a shared location from metadataserver 1. Register the table on metadataserver 2 With some coding and metadatascripting I applied the labels and formats on the table on metadataserver 2.     ... View more

I suggest you open a track with SAS Tech Support.They are in the best position to help. ... View more

I could not find the answer I was looking for. I have flows build in DI studio, when I run the generated .sas code in batch I want the batch job to stop when/if the first error occurs.   Any idea's?  ... View more

If you just want to change the semi-colons to pipes then read the file byte by byte and make the change.   If you must read the data and re-generate it then use the method I posted before to read the actual values of the data files and then add some code to write it back out in the format you need.   ... View more

wlatin1 doesn't support those characters, so even if you load the data, you wouldn't be able to work with them.  Just fire up a separate UTF8 session and do it in there, think most should have moved onto UTF8 by now anyways? ... View more

You might try adding a required prompt to display just one region (or other subset) at a time. ... View more

I have created small example for you. I hope you it goes well. data test; input x datetime.; format x datetime18.; cards; 25DEC18:00:02:09 ; run; proc sql; insert into test values ('26DEC18:00:02:09'dt); quit; proc print data=work.test; format x datetime18.; run; I also used one PROC SQL. so, it's 18 length of datetime.   Thanks --       ... View more