About BrunoMueller

BrunoMueller · ‎09-11-2017

Have a look at the CALL SORTC routine. Define an array, put your chars into the array elements and then use the CALL SORTC. What is the use case for this data manipulation?

BrunoMueller · ‎09-11-2017

Have a look at the dataStep.runCode action, if you are not reading from any table, the code gets execute on each node, this is why you get the 8 output rows. You can specify the SINGLE="NOINPUT" to run the code only one node datastep.runCode submit status=rc / single="noinput" code=sasCode ; print rc; See also DATA Step Processing Modes for more information.

BrunoMueller · ‎09-08-2017

In my example I first write out the JSON to a temp file and use the DATA Step to stream it back to the appropriate fileref as defined in the data targets of the stored process metadata. Like so: filename xjson temp; proc json out=xjson pretty; export &tableName; run; data _null_; infile xjson; input; file myjson; put _infile_; run; filename xjson clear; I do use the %STPBEGIN and %STPEND macros. I do not set the application header You can download the sample STP and Proc HTTP code to call the Web Service from the web site I mentioned earlier also note, that if your SAS code has an error, you get the code 500, also check the Stored Process Server Log

BrunoMueller · ‎09-08-2017

Hi Have you defined data targets in the Metadata of the Stored Process. By default call a STP as Web Service will return the output parameters defined as a JSON structure. To access other data, you will need to define a data target and use a different end point to call your STP as Web Service. You will find a quick introduction on using a Stored Process as a Web Service on this page https://www.sas.com/de_de/training/home/services/webinar-lunchtime.html Search for "Stored Processes als Web-/REST-Service aufrufen" You can download the slides and sample STP and SAS program code. The slides are in German, I hope this gets you going.

BrunoMueller · ‎09-05-2017

Best way would be to build a format based on the airport code, I am sure you have a data set that has airport codes and cities. You then can use the format together with %SYSFUNC and the PUTC function within a macro. Here is a simpel example proc format; value $apcode "CYVR" = "Vancouver" "CYYC" = "Calgary" ; run; %let apcode = CYVR; %put NOTE: &apcode is %sysfunc(putc(&apcode, $apcode.)); %let apcode = CYYC; %put NOTE: &apcode is %sysfunc(putc(&apcode, $apcode.));

BrunoMueller · ‎09-05-2017

You can use the POINT= options of the SET statement to randomly read an observation. To calculate the random observation number you can use the RANUNI function together with the CEIL function. The NOBS= option on the SET statement will return the total number of observations in SAS data set. The DATA step will read the data set with all the logins, you then calculate a random number to read from the password data set

BrunoMueller · ‎09-05-2017

Since Proc REPORT builds a report lione by line, we can no longer distingiush the values for the different drivetrains. To get around this, you also need the horsepower and cylinders in your across defintion, so you can do the calculation. For across columns, we can not use the varibale name, but have to use the absolute column number, including the ones that have the NOPRINT option set. See code below: proc report data=newCars; column type horsepower Cylinders driveTrain, (horsePower=hp2 cylinders=cyl2 compVar) _dummy; define type / group; define driveTrain / across; define horsepower / analysis ; define Cylinders / analysis; define compVar / computed f=commax15.; define hp2 / analysis noprint; define cyl2 / analysis noprint; define _dummy / computed noprint; compute _dummy; _c6_ = _c4_ * _c5_; _c9_ = _c7_ * _c8_; _c12_ = _c10_ * _c11_; endcomp; run; Depending on your data, this might not be the most flexible and maintenance friendly way of doing it. Maybe it is better to precalculate all the data, and use Proc REPORT for the display/layout part, see example below: proc sql; create table carsAggr as select type , driveTrain , sum(horsePower) as sumhp , sum(cylinders) as sumcyl , calculated sumhp * calculated sumcyl as mult_hp_cyl from sashelp.cars group by type , driveTrain ; quit; proc report data=carsAggr; column type sumhp sumcyl mult_hp_cyl, driveTrain; define type / group; define driveTrain / across; define sumhp / analysis f=commax15.; define sumcyl / analysis f=commax15.; define mult_hp_cyl / analysis f=commax15.; run;

BrunoMueller · ‎09-05-2017

Hi You can not use GROUP and DISPLAY together. I assume you get a note like: NOTE: Groups are not created because the usage of ..... is DISPLAY. To avoid this note, change all GROUP variables to ORDER variables. Testing this out with data all SAS users have access to to makes things easier. The following code has two examples, first using DISPLAY, so every row will be displayed, the second using ANALYSIS, where data is aggregated on the group varibale. Which one represents waht you need? proc report data=sashelp.cars(obs=20); column type horsepower Cylinders compVar, driveTrain ; define type / group; define driveTrain / across; define horsepower / display ; define Cylinders / display; define compVar / computed f=commax15.; compute compVar; compVar = horsePower * Cylinders; endcomp; run; proc report data=sashelp.cars; column type horsepower Cylinders compVar, driveTrain ; define type / group; define driveTrain / across; define horsepower / analysis ; define Cylinders / analysis; define compVar / computed f=commax15.; compute compVar; compVar = horsePower * Cylinders; endcomp; run;

BrunoMueller · ‎08-25-2017

Since the value for GROUP value is only displayed once, you have to put it aside. For this you need some logic in Proc REPORT, see sample below. However if you just need that data to be then exported, see the Proc SQL and Proc TRANSPOSE code below. proc report data=sashelp.cars; column type displayType driveTrain origin, msrp _dummy; define type / group noprint; define driveTrain / group; define displayType / computed ; define origin / across; define msrp / analysis; define _dummy / computed noprint; compute displayType / char length=32; endcomp; compute _dummy / char length=32; if type ne " " then do; tempName = type; end; _dummy = tempName; displayType = tempName; endcomp; run; proc sql; create table cars_aggr as select type, driveTrain, origin, sum(msrp) as msrp_sum from sashelp.cars group by type, driveTrain, origin order by type, driveTrain, origin ; quit; proc transpose data=cars_aggr out=cars_trsp ; by type driveTrain; var msrp_sum; id origin; run;

BrunoMueller · ‎08-25-2017

Hi You can use Proc REPORT to get what you want. For this you do need two additional computed columns. In the compute block for the _DUMMY column, we have the logic to fill the column M_NAME with the appropriate content. Have a look below. If there is not nesting between market area and market region, you could also use Proc TABULATE. data have; input market_area $ market_region $ customer_id; datalines; east A 12 east A 13 east B 10 east C 17 west D 12 west E 15 west F 18 ; proc report data=have; column market_area market_region m_name customer_id _dummy; define market_area / group noprint; define market_region / group noprint; define m_name / computed ; define customer_id / analysis n; define _dummy / computed noprint; break before market_area / summarize; rbreak after / summarize; compute m_name / char length=32; endcomp; compute _dummy / char length=32; _dummy = _break_; if lowcase(_break_) = "market_area" then do; m_name = market_area; end; if _break_ = " " then do; m_name = market_region; end; if _break_ = "_RBREAK_" then do; m_name = "Total"; end; endcomp; run; proc tabulate data=have; class market_area market_region; table market_area="Area" market_region=" " all , n ; run;

BrunoMueller · ‎08-23-2017

check the model type for the data item, should be discrete

BrunoMueller · ‎08-22-2017

Simply change the measure into a category on the data item in question. However I would recommend to have all category variables defined as character variable and likewise a measure should always be numeric, prior to loading the data into the LASR server

BrunoMueller · ‎08-18-2017

As you mentioned you already have a table with all the relevant information for all the users. You can use the Row Level Security functionality to just show the information for the user logged in. For this to work you need to set a conditional GRANT on the table in question. There are various ways of doing this, for instance using the CLI tool sas-set-metadata-access, here is an example: sas-set-metadata-access -profile SASAdmin "/Shared Data/SAS Visual Analytics/Public/LASR/USERINFO_TEST(Table)" -grant sasusers:Read -condition "userId contains 'SUB::SAS.Userid'" Another way is using the SAS Visual Analytics Adminstrator, aseen here https://communities.sas.com/t5/SAS-Visual-Analytics/SAS-Visual-Analytics-7-3-Row-Level-Security/m-p/325739#M6164 You will find more information on how to use Row Level Security in this paper: http://support.sas.com/resources/papers/proceedings16/SAS6660-2016.pdf

BrunoMueller · ‎08-15-2017

Hi You can tell Proc UNIVARIATE to use ODS Graphics to create the output. Using ODS TRACE you can figure out which output objects are created. The output objects data can be written to a SAS data set using ODS OUTPUT. Here is a code sample that illustrates the points mentioned above. It will only print the histgram plot and write all the data from the plot to a SAS data set. ods graphics on; ods trace on; ods output Histogram=histodata; proc univariate data=sashelp.cars noprint; var invoice; histogram invoice / midpoints=(0 to 180000 by 5000) ; run; ods trace off;

BrunoMueller · ‎08-15-2017

Hi You have to tell Proc FREQ to actually use the formats you have just created. A format definition is an indepent element and can be used with many different tables. The format name does not have to be the same as the column name. The FORMAT statement makes the association between a variable and the format to use. Example: PROC FREQ data= WORK.JOINED6; tables arrival_fulln*work_status_fulln / chisq; FORMAT arrival_fulln arrival_fulln. work_status_fulln work_status_fulln. ; run;

Online Status	Offline
Date Last Visited	4 weeks ago