Partial least squares logistic regression is a good method of dealing with collinearity in logistic regression. Currently, it is not supported by SAS, so I think it necessary to include it in SAS/STAT. ... View more

I think having an extension for visual studio code would be amazing. It would add another avenue for SAS Users to integrate their code into repo like github and gitlab, which in turns allows better versioning control for teams and pipelines.   Features Highlight incorrect syntax such as ';' missing autocomplete feature for procs and commonly used syntax color coding for specific things like proc and such.  suggestive typing     This would function similar  https://blogs.sas.com/content/sasdummy/2017/08/25/npp-with-sas/ but for visual studio code.  Someone has started (https://marketplace.visualstudio.com/items?itemName=77qingliu.sas-syntax) but it still missing a lot of support and it something that SAS could help with.       -Regards Jake ... View more

When using SET datasetname POINT= idnumber   the user will create an infinite loop, unless there is a STOP; statement that is executed. It is easy to forget this, or "hide it in some logics.      Of course an experienced SAS user NEVER markes this mistake - until the other day.... Suggestion: Try to implement a check that there is a STOP statement, in the SAS Compiler. / Br Anders Sköllermo ... View more

Hi! Proc Means and Proc Univariate at present lack fast and accurate methods for calculation of quantile, means and standard deviation for large sets of data. In the paper "Fast and Accurate calculation... "  by Anders Sköllermo I point at a possible method.   Suggestion: Look at this method. Develop it further to be a general method and implement it in the SAS procedures. Future: I think that this method can also been very good when using several computers or parallel computers or computer networks   and also on streaming data.  However the method has not at al been tested for this. (The environment I had was SAS on demand for Academics). /Br Anders Sköllermo ... View more

"Start Page" currently contains two blocks of items in SAS Enterprise Guide - "Pinned Items" and "Recent Items".   The "Recent Items" section should be split into "Recent Programs" and "Recent Datasets" based on the type of item being referred to. Each section should be selectable for "viewing" either in the "Start Page" (e.g. grouping drop down) or in SAS Enterprise Guide's menu (Tools > Optons") so that they are hidden or visible. Further sections could be added, so that file paths (and SAS logical references) are used to determine the content "relationship" to SAS - .sas for "programs", .csv/.xls/.xlsx etc for "files" (or "data"), etc. ... View more

The SAS EG editor, where I spend most of my day, could be much improved. I'd like to make the following suggestions to make my work easier and more efficient:   When opening a file in SAS EG, it doesn't use a standard Windows file open dialog box. Instead, you must first select from Recent, My Computer, Servers, etc. If you select My Computer or Servers, you then need to select the folder. It should just open to the last accessed folder with one click. Making all of these selections every time is annoying and unproductive. Because it does not use a standard Windows file dialog, the folder of the file you are editing is not displayed, but it should be. Since you have the ability to open local or remote files, it would be very helpful if you had a window similar to WinSCP, enabling easy file transfer between your local machine and a server. In the log window, clicking on Errors, Warnings, or Notes does not correctly subset the list below to display Errors, Warnings, or Notes. Block copy (Alt + left mouse button + drag) does not work correctly (as it does in Notepad++). You can't create a block past the end of a line - you need to insert spaces first, or you will have a jagged left edge to your selection. SAS should automatically insert spaces to match the width of the selection or the line above. You can't select a vertical line and begin inserting characters. If you want to insert a vertical block of text, you need to insert it one row at a time manually. This functionality is very easy in most editors and would save me a lot of time. When performing a replace in a selection of text, the replace dialog should stay open and the selection stay selected so that you can make another replacement without having to re-select the text. It would be nice if you could easily flag a row by clicking in the left margin, similar to what you can do in Notepad++, so that you can easily find that row again after scrolling up or down. It would be helpful to have a count of the number of rows selected at the bottom of the screen, like in Excel, so that you can easily count code lists and the like. ... View more

As I understand the situation, a "spike and slab" prior in PROC MCMC cannot now be specified because this distribution requires a combination of a DGENERAL and GENERAL distribution. [Regarding "spike and slab", see the reference:  https://www.mv.helsinki.fi/home/mjxpirin/HDS_course/material/HDS8_variable_selection.html] The "spike and slab" prior for a predictor X would support "variable selection" when using PROC MCMC.  To fully implement the approach given in the reference above, the functionality of the R package called BoomSpikeSlab would be needed. See https://cran.r-project.org/web/packages/BoomSpikeSlab/index.html ... View more

Perhaps due to my ignorance, but I had a use case where WHICHC function was ideal. The problem was that the array I had was multidimensional, so I had to resort to DO i=1 TO...looping. Just allowing a multidimensional array to be specified, but with 2nd dimension specified with a constant would be an improvement: do i=1 to x; pos = whichc('#',of space[*,i]); end; ... View more

Sample code, based on SAS Usage note 22883 http://support.sas.com/kb/22/883.html   data a; do Rater1=1 to 2; do Rater2=1 to 2; input Wt @@; output; end; end; datalines; 0 3 0 3 ; run; proc freq data=a; weight Wt / zeros; tables Rater1*Rater2 / agree nocol norow nopercent plots=none; run;   This code generates a 2x2 table to estimate Cohen's Kappa with standard error and confidence interval. One of the margins has a sum of zero, but it is forced to fit a 2x2 table using the weight statement and zeros option. The results are that the estimated Kappa is zero, the standard error is zero and the Confidence Interval is (0 to 0).   I've convinced myself that the standard error result is numerically correct, having checked the result using equations in the documentation and the cited reference. I think it is a special case for 2x2 tables, since the usage note above gives an example for 4x4 tables, and the resulting standard error is strictly positive. That said, an estimated statistic having no variability might not appear correct to a reader.   I suggest that 1) PROC FREQ provides a warning/note in the log/output to draw attention to this anomaly. 2) Ideally, if there are suitable improved formulae that don't have this defect, could they be implemented as an alternative option? (I admit, I don't have a better suggestion in mind).   Reference: Fleiss JL, Cohen J, Everitt B. (1969) Large sample standard errors of kappa and weighted kappa. Psychological Bulletin. 72 (5) 323-327. https://api.semanticscholar.org/CorpusID:85541244, https://doi.org/10.1037/H0028106 ... View more

Currently there is no attribute setting to set a fill pattern for bar charts within the attribute map data sets.  It would be helpful to be able to add pattern settings using a discrete attribute maps.  https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/grstatgraph/n1oq0axfb1tc1yn10kgw6dyuitk2.htm   It would be helpful to have a FillPattern as a column name.   (PS. I know you can get around this using GraphData via template + FillStyleELement)   ... View more

Hello! I would suggest to integrate a file diff functionality in EG. Currently EG relies on external tools that need to be configured ... this is fine... but in certain companys it is just wimmering and wineing you hear when you ask for some 3rd party software. So something simple that ships with EG would be just great! 🙂 Cheers fja  ... View more

To view how well the model discriminates those with and those without the event, the user needs the area under the receiver operating characteristic curve (AUC). However, the options for the users are limited when the INMODEL= option is specified in the PROC LOGISTIC statement, with AUC only available when the user specifies both the SCORE statement and the FITSTAT option. But a big deficit in the process is the lack of confidence interval of AUC, which is automatically computed and displayed in the ROC statement but is not available when the INMODEL= option is specified in the PROC LOGISTIC statement. Therefore, when the user is validating a model, he/she cannot obtain it directly. What is more, the formula of the standard error of the AUC is not reported in anywhere of SAS Help, so the user is unable to calculate the confidence interval on his/her own. Therefore, I suggest that SAS (1) automatically compute the standard error and confidence interval of the AUC when the INMODEL= option is specified in the PROC LOGISTIC statement; (2) include the formula of the standard error of AUC in SAS Help. ... View more

Hello, im suggesting that you implement some sort of overview of which variables that are used in a certain database for a SAS VA report.   For an example if you using a database with the paramaters a,b,c,d,e,f and are only using a,b,c, then a list that contains the used variables would be perfect (in this case only a,b,c). At the moment you have to go in on a report, click up the data and hover the mouse over each variable to see if it is used or not.  ... View more

The OUTSTATS dataset created by PROC COMPARE includes summary statistics for only numeric columns, eg. N, MEAN, STD, NDIF.  While the MEAN, MIN, MAX, STD, STDERR etc values are not applicable to character columns, N and NDIF are.   The OUTSTATS data is very helpful to very quickly get a sense of the magnitude of differences between the BASE and COMPARE datasets.  e.g.   outstats = column_diffs ( where = ( _type_ = 'NDIF' and _base_ ne 0 )). The ODS table output produced by "ods table comparesummary = blah" is just the printed output that must be parsed, an untenable solution.   Please add N and NDIF to the OUTSTATS data for character columns. ... View more

I've noticed that when i import a DI job that uses either a SQL execute or a user written code transform. The user written code that is imported to VIYA is on a single line and you must scroll to the right to see the code. It would be useful to either auto format the code upon import or provide a auto-format button. At the moment my work around is to copy the single line of code and then create a new .sas program, paste the code in and then auto-format. I then have to paste my code back into the node within the swinlane and then save. On the screen shot below, my original code has 100s of lines of code however it is imported onto a single line.   The .spk is a DI job. We are on a 9.4 m6 AIX 7.1 deployment... not sure if this is anything to do with the operating system. ... View more

Currently, PROC MIXED can implement just one type of heteroscedasticity consistent estimator of the variance, namely the Huber-White / empirical / sandwich / HC0 estimator; this is accomplished with the EMPIRICAL option on the PROC MIXED statement in conjunction with the SUBJECT= option on either the RANDOM or REPEATED statement:  https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.4/statug/statug_mixed_syntax01.htm#statug.mixed.procstmt_empirical   PROC REG, on the other hand, can implement 4 types of heteroscedasticity consistent estimators: HC0, HC1, HC2, and HC3 using the HCC option in the MODEL statement: https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.4/statug/statug_reg_syntax08.htm#statug.reg.modelhcc   Long and Ervin (2000) argue that HC3 is superior to HC0.   Additional methods such as HC4, HC4m, and HC5 have been proposed in the literature and implemented in the vcovHC function of R package sandwich.    Please consider adding these additional heteroscedasticity consistent estimators to PROC MIXED.   Relevant references:  Long J. S., Ervin L. H. (2000). “Using Heteroscedasticity Consistent Standard Errors in the Linear Regression Model.” The American Statistician, 54, 217–224. Cribari-Neto F. (2004). “Asymptotic Inference under Heteroskedasticity of Unknown Form.” Comuputational Statistics & Data Analysis 45, 215–233. Cribari-Neto F., Souza T.C., Vasconcellos, K.L.P. (2007). “Inference under Heteroskedasticity and Leveraged Data.” Communications in Statistics – Theory and Methods, 36, 1877–1888. Errata: 37, 3329–3330, 2008. Cribari-Neto F., Da Silva W.B. (2011). “A New Heteroskedasticity-Consistent Covariance Matrix Estimator for the Linear Regression Model.” Advances in Statistical Analysis, 95(2), 129–146. ... View more

For debugger we can sort based on variable and based on value, but there is no way to sort for WATCH variable. an there be a functionality added that will allow us to filter on the WATCH variables so they appear together versus where they appear in the PDV? ... View more

Hi,   I'm suggesting if you could let the user add two legends to the Line-Bar chart as below screenshot: to show more details to the info based on two measures and two categories    ... View more

While many welcome the availability of SAS EG for connecting to SAS Viya 4 as they see it as a continuity from the past (and carry the historical baggage along with), the availability of multiple interfaces to connect to SAS Viya 4 (SAS EG, SAS Studio, VS Code ) is  chaotic and an avoidable irritant for admins.  In my experience VS Code has proved it's mettle . When it works for   Java, C#, go, python and, C++,   Oracle etc.,  it can work for SAS too as efficiently. It is my humble suggestion that SAS promote VS Code as the official Code editor for SAS/CAS programming and deprecate SAS Studio and SAS EG over a reasonable period of time. This would simplify managing the environment (and make business sense to SAS too!). SAS would only need to focus on extensions. ... View more