Hi:   You might want to go back and revisit the lectures on the RETAIN statement and the SUM statement. In this code (I changed the output data table names with 1 and 2), You can see that the only difference is the RETAIN statement. The fact that Example 2 works the same as Example 1 tells you something about how the SUM statement works.   Notice that the lecture on the SUM statement explains that the SUM statement does an automatic RETAIN.   Thus, the RETAIN statement in the first example is not necessary for purposes of having the program work. On the other hand, using a RETAIN statement, even with a SUM statement makes it clearer that totTraffic is the variable being accumulated and retained.    Understanding when you need to use RETAIN and when it's not absolutely required is one of the objectives for this practice. However, you can break the SUM statement like this:   Learning when to use the RETAIN statement and how the SUM statement works is one of the goals of this Lesson. Cynthia   ... View more

Hi,   Thanks for explaining what you're using to study. The data for some of our free classes, like Programming 1 and Statistics 1 is available from inside the free e-learning version of the class. However, the Machine Learning class is not one of the free classes.   The Machine Learning class data is provided only on our Virtual Lab images. When a student purchases a Learning Subscription with the Machine Learning class or when they take the class as an academic user then the data is pre-loaded on the lab they use for class.   So that is why you're not finding the COMMSDATA table when searching for publicly available data for the class. You can look for some alternative data in the SAS documentation. Sample data for SAS Viya platform: https://support.sas.com/documentation/onlinedoc/viya/examples.htm Sample data for SAS Visual Analytics: https://go.documentation.sas.com/doc/en/vacdc/v_031/vareports/n1dj9jn3klf501n1ovpkkahcegot.htm My understanding is that you should find the HMEQ data as publicly available data to use to practice. Here's a video that shows using HMEQ data in Model Studio: Demo using HMEQ Sample data: https://www.youtube.com/watch?v=Ep3p7WzptCw Cynthia ... View more

Hi, There are several different ways that we provide a SAS Viya environment to practice the Machine Learning course: 1) SAS  Viya for Learners which is the platform available to academic users to practice; 2) Virtual Lab inside the Machine Learning class that is available with a paid subscription; 3) Virtual Lab interface accessible from inside Coursera/SAS classes.   However, the class name you say you are taking or reading sounds more like one of the classes on our platform, not on the Coursera platform.   Depending on which of these methods you are using, you should find the COMMSDATA file when you are in Visual Analytics or in Model Studio. However, if you are using SAS  Viya at your company installation of SAS Viya or in one of our Viya free trial packages, our class data would not be loaded at your company and our class data is not loaded in the Viya free trials.   When you say that you've looked in SAS Studio for the data, depending on which SAS Studio you're using, you would not find the data in SAS Studio on your work installation of SAS or  on the SAS OnDemand for Academics server. If you are a university student and are using SAS Viya for Learners, then you should be able to find the data when you use Model Studio or Visual Analytics or SAS Studio under the Develop Code and Flows menu.    Can you clarify how you're using SAS Studio and how you're taking the course? For example, did you sign up for the course on the SAS web site? Or on the Coursera web site? Or is your University professor teaching Machine Learning using SAS Viya as a  text in your class--in this case, I would expect your professor to also be having students do practices on the SAS Viya for Learners platform.   Once we know better which SAS Studio you're trying to use to  open the COMMSDATA file, it will be easier to help you. As I remember most of the practices in the Machine Learning class use SAS Viya and Model Studio. The first demo in the most current version of the Machine Learning class has students open Model Studio and navigate to the COMMSDATA file, which is already loaded on the SAS Viya for Learners server. That first demo does not have students open COMMSDATA with SAS Studio.   What version of the class are you taking?  If you are using SAS Course Notes, then most of our courses with course notes assume that students are using the classroom image that is loaded in the Virtual Lab. University students using SAS Viya for Learners have slightly modified instructions. Knowing how you are using SAS  Studio, which platform and which version of the class would be very useful information. Cynthia ... View more

Hi:   That is not what I observed. Did you run my test program using my fake test data? With the test data and nearly the same program (columns=2 instead of columns=3) that specific character string worked in the WHERE statement in my test program. If the test program works on my test data, but the WHERE does not work on your real data, then I would suspect some issue with the way the values are stored in the data set....either unprintable spaces, two underscores or two hyphens instead of 1, leading spaces....something is off with the data.   If you run my test program with MY fake test data, and the WHERE in the SGPANEL does not create the graph that I pasted, then, that is an issue you need to address with Tech Support.   My test data and test program proves that the string '464918_99' can be used in a WHERE statement and the condition works correctly, when '464918_99' is in the data.    You have not provided any sample or representative data or shown a SAS log which should have something like THIS after the SGPANEL step that shows EXACTLY how many observations were selected based on the WHERE:  You keep saying that the character string is not being 'read' by the WHERE statement, which is clearly NOT the case in my test program. Just saying over and over that the value is not being read doesn't help anyone make constructive suggestions because you haven't said whether Variety is a character or numeric variable in your data; you haven't said whether Variety has a user-defined format in use; you haven't shown any results or log messages that illustrate your assertion that the value is not being read or used by the WHERE statement. On the other hand, several community members have made suggestions and provided code and asked questions that all point to other things you need to check before you can really say that the WHERE statement is not 'reading' the value '464918_99'.   So, did you run my ENTIRE test program and what EXACTLY did the log show and what EXACTLY were the results? Cynthia     ... View more

Hi:   A character string should be OK whether it is all numbers or a mix of numbers and characters or just all alpha letters. What might make a difference is leading blanks or unprintable characters or a hyphen - in the value instead of an underscore _.   These are the results from my test:   I made some fake data from SASHELP.CLASS and just made up fake values for PeelColor. I made data for 4 seasons, but one of the seasons didn't pass the WHERE and I had one season where the varieties didn't pass the WHERE, so the resulting SGPANEL only had 2 Seasons but did correctly show the value 464918_99 as a character value.   Here's the data and code I tested with.  Cynthia ** Make some fake data using SASHELP.CLASS; ** Data for 4 season values; ** One Season is not wanted in the WHERE; ** One Season has bad varieties; ** final SGPANEL should only show 2 seasons; ** if WHERE is working correctly; data one; length variety $15 PeelColor $4; set sashelp.class; ** final data should be good for 2021 and 2024 with WHERE; ** Seasons 2022 and 2023 should not pass WHERE; Season = 2021; if sex = 'M' then variety = '464918_99'; else if sex = 'F' then variety = 'Hass'; if age le 13 then PeelColor = 'Lite'; else if age le 16 then PeelColor = 'Dark'; output; ** SEASON 2022 is OK, but both Varieties are not OK; Season = 2022; if sex = 'M' then variety = 'Unknown'; else if sex = 'F' then variety = '589333_44'; if age le 14 then PeelColor = 'Lite'; else if age le 16 then PeelColor = 'Dark'; output; ** Season 2023 should fail where because SEASON is wrong; Season = 2023; if sex = 'M' then variety = '464918_99'; else if sex = 'F' then variety = 'Hass'; if age le 14 then PeelColor = 'Lite'; else if age le 16 then PeelColor = 'Dark'; output; Season = 2024; if sex = 'M' then variety = '464918_99'; else if sex = 'F' then variety = 'Hass'; if age le 12 then PeelColor = 'Lite'; else if age le 16 then PeelColor = 'Dark'; output; run; title "Ripe PeelColor: 464918_9 vs Hass by Season"; proc sgpanel data=one pctlevel=group; where Season in(2021, 2022, 2024) and Variety in('464918_99', 'Hass'); panelby Season / columns=2 onepanel; vbar Variety / group=PeelColor grouporder=ascending stat=percent groupdisplay=stack seglabel; run; title; ... View more

Hi: I still don't understand what is wrong with the way your merge is working. Your PDF examples do NOT have any errors and you do not show any wrong results, correct results or explain what results you want. Cynthia ... View more

Hi:   As several members have pointed out, your original code runs without error. Therefore, you either have a data problem or a program logic problem. Without more definition from you about exactly why you think your merge is not working, with examples, and what you would consider to be the correct results, with examples, it is nearly impossible to provide constructive suggestions. Several of us have posted information about how your original code is working without errors. There have also been repeated guesses about whether your using the correct variables in the merge or whether you need to merge by both ID and NAME or whether there is something wrong with your data. However, without some further clarification from you, it is hard to help. Cynthia ... View more

Hi:  In order to test the "_NEWC_" variable inside your PROC REPORT code, it MUST be on the COLUMN statement for PROC REPORT. Even if you have the variable in your data, PROC REPORT has to know about the variable before you can use it in an IF statement in a COMPUTE block. Then, as indicated, the placement of the variable on the COLUMN statement makes a difference. So if you are going to test the "_NEWC_" variable in the COMPUTE block for RCVD then it has to appear BEFORE RCVD in the COLUMN statement. There are other methods that may work too, but they are more complex. Cynthia ... View more

Hi:   If you run both programs, as suggested, you'll see that the results of both programs are the same whether you use a WHILE or an UNTIL:       However, look at the condition that was needed to make each DO loop work to produce the same results. The UNTIL code uses (Capital GT 500000) The UNTIL will ALWAYS have the loop executed at least once because the condition is checked at the bottom of the DO loop, after the statements have executed at least one time.   The condition for the WHILE was (Capital LE 500000) and the condition for a WHILE is checked at the top of the loop. So in this example, both loops would run the same number of times to produce the same results.   If you have data and some logic that needs to be executed at least one time if the initial condition is met, then use a DO UNTIL. If you have some logic that you do NOT want to execute if the initial condition is met, then use a DO WHILE. This first example shows that both forms of the DO loop are capable of producing the same results, given that you have coded the condition correctly.   However, to prove that it is possible for the 2 loops to produce different results, what if in the example program, we started the value for Capital at exactly 500000 and what if we changed the logical operator in the condition for the WHILE -- so that instead of testing for LE, we just tested for LT -- then the results WOULD be different:   Now you see that the program with DO UNTIL produced different results than the program with DO WHILE. This is not always the case, as the first program proves. You are in control of the operators used for the DO loop conditions.   The important things to learn about DO UNTIL/DO WHILE are:  where does the condition for the loop get tested? Then, for the data and problem you are working with, do you need for the statements to execute at least one time or not? Finally, what logical operator do you need to use to make your loop work with your data? In the first program, we wanted to illustrate how it was possible for both forms of the DO loop to produce the same results. The key to making that happen was the logical operator we used in the code for the question.   Hope this helps explain things a bit more. Cynthia ... View more

Hi:   I'm just trying to understand the logic of how Key_ID 3 has a value of 31 for Want_January when Date_1 is in Dec 2023 and how Key_ID 4 has a value of 22, again, when the Date_1 value is in Dec 2023. Your logic is a bit murky. Can you clarify? Cynthia ... View more

Hi,   When I run a slightly modified version of your code, I do not observe the same behavior. Here's my output:   And here's my code: options label orientation=portrait leftmargin=.10in rightmargin=.10in; ods noproctitle; ods pdf file='c:\temp\checkODS.pdf' style=journal startpage=no; ods layout gridded columns=2 advance=table; ods region; proc freq data=sashelp.class; title 'SASHELP.CLASS'; table sex*age/nopercent nocum norow nocol; run; ods region; proc freq data=sashelp.classfit; title 'SASHELP.CLASSFIT'; table sex*age/nopercent nocum norow nocol; run; ods layout end; ods pdf close;   So I'm not sure what is happening to cause your code to get a blank page. You might want to work with Tech Support on this.     Personally, I don't like the look of PROC FREQ output side by side and I probably would have used TABULATE or REPORT instead.   Cynthia ... View more

Hi, My experience using images with PROC REPORT and inserting the image into a cell is that the image does not "autosize" -- so I generally had to use a paint program to get the image to the approximate size I needed before I inserted it into the output. It may also make a difference whether your ODS output destination is HTML, RTF, PDF, etc. Cynthia ... View more