snoopy369 Tracker

Re: Discussion，AI [Agent] with SAS?!

snoopy369 — Wed, 29 Apr 2026 13:55:00 GMT

Check out my presentation from SAS Innovate yesterday - slides here: https://github.com/matise-joe-norc/connect-saspy

I cover using Claude Code with SAS!

Re: What is the state of SASGF 2022?

snoopy369 — Tue, 26 Jul 2022 17:21:48 GMT

Another option that I've seen is co-authoring the paper with someone who is attending the conference. I've presented papers on behalf of people outside the country who couldn't attend, as have others I've worked with in the past. Usually there will be someone who is familiar with the topic and is happy to collaborate, even if the collaboration is 90% presenting it and 10% talking it over.

Re: Installing add-on package

snoopy369 — Tue, 14 Jun 2022 22:32:02 GMT

The instructions in the link are correct regardless of the IDE you're using, EG or DM. If you're running EG with a local install, it's just calling sas.exe the same as DM would be if you ran it directly. All you do is copy the .dll files to a particular folder. (You can also copy them to a different folder and add that folder to the PATH in your sas config, if you know how to do that.)

If you don't have a local install, then you can't install PROC TRAJ directly, you have to talk to your server admin and they have to install it on the server. It's the same concept (put it in the right folder and/or in another folder in the PATH), but it's on the server not on your computer.

Re: Help with ods output statement

snoopy369 — Mon, 21 Feb 2022 23:55:47 GMT

Not an expert at regression process, but you could try the PERSIST option (see https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/odsug/p0oxrbinw6fjuwn1x23qam6dntyd.htm) to see if it will keep it open as opposed to overwriting it.

Re: drill down from visual SAS Visual analytics

snoopy369 — Mon, 14 Feb 2022 18:01:05 GMT

This is done using linking.

First, you create a second page, which includes *all* of your detail data. Here, I use a separate dataset that has the detail data; you could do this entire report with just one dataset if you had a computed total, but this is slightly easier - I like to do summaries outside of VA.

data htemp.bike_usage;
input month miles;
datalines;
1 100
2 200
3 300
4 400
5 500
6 600
7 500
8 500
9 800
10 1200
11 1100
12 1200
;;;;
run;

data htemp.month_detail;
input month means_of_transport $ miles;
datalines;
9 Bicycle 200
9 Bus 200
9 Car 200
9 Van 200
10 Bicycle 300
10 Bus 400
10 Car 200
10 Van 300
;;;;
run;

Then load these to CAS if you haven't already, and then I make a report, using the month summary dataset for the chart you made (the bar chart), titled "Overview", and then a second tab, "Detail", which is a list table. You can of course do something other than list table, but the important thing is that it needs to have "month" as a field somewhere - it can be a "hidden" field, but it has to be associated with the chart somehow.

Then, go to the Overview report and select Actions, then the bar chart, and open Page Links. There should be a selection for Detail, and a note that you need to map the data (since it's from a different dataset).

Click on Map, and if the month variable is the same name in both datasets it'll auto-select as the mapping criteria - if not, pick the right mapping variable (think the variable which tells SAS which row should be mapped to which row from the left and right datasets).

Once you've mapped the variable, then bars should become clickable, and you'll get the drilldown you wanted! It will filter the Detail tab with the Month value from the bar. Note the breadcrumb at the top which lets you go back easily, and the filter button showing what was filtered by.

Applying control-based filters to subset of data

snoopy369 — Mon, 14 Feb 2022 17:00:22 GMT

**This is for SAS VA 8.5.2 in a Viya 3.5 environment.**

I am building a map that we are going to use to assist managers in assigning cases to interviewers, and running into a challenge; wondering if SAS VA can support this workflow, or if it's not feasible.

We have two source datasets, Interviewers and Respondents. Both have lat/long for each, a variable for a strata (interviewers of one strata work respondents of same strata), and some other variables unique to the individual dataset. In the example below, those are represented by Available and Assigned; in the real data there are others. These variables cannot be combined into a single variable shared across the datasets (even though in the example data, that would probably be the easiest solution).

What options do we have for enabling showing, for example, all of the interviewers with only respondents selected "Assigned=0"? Or all of the respondents with only interviewers selected "Available=1"? Or both? For this purpose, the only solution I've found so far is switching from the button bar to the list (which allows multiple select), but that's not ideal, both because one of our actual filters has one option per record (interviewer name), and it requires the user to intentionally select the "missing" option. (Of course, we would use a drop-down list not a button bar in that case - using button bars here to make the example easier to see in a picture.) We also really like the idea of the drop down list due to the ability to type in and get a name autofilled - but that is a single select option, which would be fine if it also allowed missing.

I've thought some about trying to keep the datasets second (not using the "combined" dataset but using both datasets separately), but I don't see that it's possible to do that on a single map, even if I define a relationship between the fields.

Thanks!

Example code:

data fimap.example_interviewers;
  set sashelp.zipcode;
  length type $11;
  keep x y type strata available;
  where mod(zip,580) = 0;
  type='Interviewer';
  strata = mod(_n_,8) + 1;
  available = mod(mod(_n_,3)+1,2);
run;

data fimap.example_respondents;
  set sashelp.zipcode;
  length type $11;
  keep x y type strata assigned;
  where mod(zip,53) = 0 and mod(zip,580) ne 0;
  call streaminit(7);
  strata = rand('Integer',1,8);
  type='Respondent';
  assigned = mod(_n_,10) gt 0;
run;

data fimap.example_combined;
  set fimap.example_interviewers fimap.example_respondents;
run;

Example of what the map might look like:

Re: EG 7.1 code lost during update

snoopy369 — Thu, 16 Dec 2021 19:58:51 GMT

I have had that issue back when I used 7.1. Best advice I have is upgrade to 8... doesn't seem to have that issue anymore.

Re: Is there a way to use FIRST and LAST options within an IF THEN statement?

snoopy369 — Thu, 07 Oct 2021 20:09:48 GMT

Remember that the data step is row-by-row always, it (normally) can't seek about the dataset randomly. (There is a way to do that, but it's nearly always the wrong approach - it's very slow and loses a bunch of functionality.)

What `first.var` does is it tells you whether the row you're on is *different from the previous row*. That's it! And what `last.var` does is it tells you if the row you're on is *different from the next row*. (Yes, SAS does sort of peek at the next row for this, it's the exception to the rule.) Mentally, in english, it tells you if you're on the "first" row of a sequence of that ID. It just returns 1 or 0 - true or false.

Unfortunately, it only works with the `by` variables - so we need to make your data work for that. We do need to subset the data, but we don't need to do it in a separate step - we can do it on the fly.

1. Let's subset your dataset while we do this, using where to only include the Phase1 folks.

where findw(VISIT, 'PHASE#1') > 0;

2. Let's retain the start variables, because we'll need them to hang around until the last row.

retain trtsdt trtstm;

3. Let's save aside the trtsdt and trtstm when we are on a first.id row.

if first.id then do; trtsdt = datepart(stdtc); trtstm = timepart(stdtc); end;

4. Let's then save the trtedt/trtetm when we're on a last.id row, and output that row.

if last.id then do; trtedt = datepart(stdtc); trtetm = timepart(stdtc); output; end;

All together we have:

data want;
  set have;
  where findw(VISIT, 'PHASE#1') > 0;
  by id;
  retain trtsdt trtstm;
  if first.id then do;
    trtsdt = datepart(stdtc);
    trtstm = timepart(stdtc);
  end;
  if last.id then do;
    trtedt = datepart(stdtc);
    trtetm = timepart(stdtc);
    output;
  end;
  keep id trtsdt trtstm trtedt trtetm;
run;

Re: Conexão SAS com Python para executar uma rotina

snoopy369 — Thu, 07 Oct 2021 18:11:59 GMT

@ErosGuedes wrote:

I'm starting now in SAS, I have a python robot that makes a copy of a file to the SAS server. I can make the connection via scp, and I make the copy this is ok.

I wanted via python to send a submit to run a SAS project on the server, but I'm not managing to make the connection. The Server uses SAS Guide, my machine has python 3.8.

I've read a lot of things, but I don't quite understand what I need to do to get the connection. Can anyone explain to me how to create this connection?

(Translated via Babelfish)

Are you using SASPy to connect to SAS 9, or are you using Viya?

SASPy has good documentation here: https://sassoftware.github.io/saspy/

Viya is a bit different, but for the most part easier, you can look at the documentation here: https://blogs.sas.com/content/sgf/2020/04/15/getting-started-with-python-integration-to-sas-viya-part-1-making-a-connection/

If this doesn't help, I recommend posting some more information - what kind of connection are you making? What library are you using? How do you log in to the SAS server (username/password, IOM, SSH, etc.)? What kind of server is it (linux/windows/etc.)? What kind of machine are you running Python on (linux/windows/etc.)?

Re: How to transpose many columns to one

snoopy369 — Thu, 07 Oct 2021 17:20:47 GMT

You should just do this in the data step - easier and faster than coercing PROC TRANSPOSE to do what you want! Just make an array of your variables and then `output` new rows when you have a value.


data want;
  set have;
  array recs record1-record3;
  do _i = 1 to dim(recs);
    if not missing(recs[_i]) then do;
    	record = recs[_i];
    	output;
    end;
  end;
  keep record id;
run;

Re: Using a SAS Dataset instead of an External file in this interesting Example

snoopy369 — Thu, 07 Oct 2021 17:13:48 GMT

Sure , you can do basically the same thing not using an external file. Your approach seems reasonable to me.

if numYieldConstraint gt 0 then
do;
   do i = 1 to numYieldConstraint;
       outputvalue=scan(Yield_Constraint,i,' ');
       output;
    end;
end;
keep outputvalue (... other things ...);

Using the `output` statement you can output multiple rows. You might want to output to a different dataset if you also want a one-row-per dataset (the FarmerAnalysis dataset you have now).

SAS VA error with geography fields

snoopy369 — Tue, 05 Oct 2021 15:23:12 GMT

I have a SAS VA 7.5 installation (Windows) and have run into an odd issue.

When a user selects to create a new geography field, two things happen:

They get the red generic error message, which leads to a log entry that suggests there was a CrossDomainProxy error
They are only able to make a geography from longitude and latitude, nothing else (not from zipcode or county or etc.)

The odd thing is that the map itself loads fine with lat/long geographies! I've seen the CrossDomainProxy error before (and searching online, many others have) when the OSM maps can't be loaded properly, but we don't have that issue; there's no proxy it needs to use, just a firewall.

I'm guessing there is some issue with the program failing to get to an online dataset that it uses to look up coordinates, but I don't see where I can adjust this setting. In a prior install of SAS Visual Analytics 7.4 on another machine, we had no issue with this. I had assumed it was using SASHELP.ZIPCODE to look up zipcode centroids, not something online... so I'm further confused.

Does this suggest a firewall issue, and if so, how can I identify the issue and what port/etc. to open? Or is there a configuration setting wrong somewhere - like, does VA not know where SASHELP.ZIPCODE is and is trying to look somewhere that's causing this error (or maybe it's just an unrelated error...)?

Thanks!

SAS Add-in for Microsoft Office 8.3, deploying with a shared profile

snoopy369 — Fri, 27 Aug 2021 22:36:37 GMT

I'm setting up our Add-In for MS Office installer, and want to deploy with a shared profile (the SDSControl.xml that we use from EG). It looks like it's possible to do that - but the documentation is shared between EG and the AMO ("The SAS Enterprise Guide and SAS Add-In for Microsoft Office clients update their profile (manually or automatically) from a shared profile stored in a central location"), and only mentions EG in the actual instructions.

https://documentation.sas.com/doc/en/bicdc/9.4/bidaag/p0yy6siuqppzmon1vf1ghpyw43kw.htm#n0foluzdvezttqn1bljw9n3qz6ws

Can I put the SDSControl.xml file in the Add-in for MS Office product folder, or does it go in a (not extant) Enterprise Guide folder in the depot ( \\depot-host\Depot_date_type\products\eguide__release_cd915__win__en__sp0__1) or somewhere else? And if it is in the EG depot, does it matter what release I pick?

Or is this not possible for the add-in - or does it need to be somewhere else?

Re: User Logons for Office Analytics Windows 9.4m6 platform

snoopy369 — Fri, 27 Aug 2021 22:30:52 GMT

Maybe you don't have SAS EVM installed (or installed properly), it's not mandatory for SAS 9.4.

If nothing else you could parse your Metadata Server Logs... that contains information about everybody who logs in and, I think, enough info to find out the client.

We also have a separate program, ESM (Enterprise Session Monitor - https://boemskats.com/products/esm/) which lets us do things like this a bit better than EVM does. They've actually been acquired by SAS, so it might be easy to get their stuff added on.

Re: User Logons for Office Analytics Windows 9.4m6 platform

snoopy369 — Fri, 27 Aug 2021 22:12:18 GMT

Can you get what information you want from the SAS Environment Manager reports (https://go.documentation.sas.com/api/collections/evcdc/2.5_M1/docsets/evug/content/evug.pdf?locale=en - see around page 92-95)

Re: Options for Onedrive integration with SAS Studio/EG

snoopy369 — Fri, 27 Aug 2021 20:19:42 GMT

Oh, ok, you're talking about stand-alone installs - that's what I'm trying to avoid 🙂 And I'm the SAS Admin, am talking with the IT folks, just hoping some other SAS Admin types might have ideas for how it's done elsewhere.

Re: Options for Onedrive integration with SAS Studio/EG

snoopy369 — Fri, 27 Aug 2021 17:48:06 GMT

I am confused ... maybe there are some mistaken assumptions on my part here. What does this mean... " all that should be needed to would the common libraries they use to be mapped from the server."? For network shares (say, \\network\share\project\subfolder\file.xlsx) then yes, that's what we do now - but they have files in OneDrive or in Teams that aren't on the network shared drives.

When you say the OneDrive path is c:\...\whatever\ is that on the server's c:\ ? That's what the path is on our local laptop, but the server can't see the laptop of course. Did your IT set things up so that they (also) have a "synced" onedrive folder on the SAS server? Can we do that without having them log in interactively? And I notice that there's two folders on my laptop - one for _my_ onedrive, and one for my "teams" folders - maybe I have a misconfiguration, but we'd need to figure both out, of course...

Thanks!

Re: Backup codes SAS VIYA 3.4

snoopy369 — Fri, 27 Aug 2021 16:40:53 GMT

One great option is SAS Studio integration with Git.

Re: SAS Web Apps routed to 8343 instead of 443

snoopy369 — Fri, 27 Aug 2021 16:38:41 GMT

Windows or Unix backend? 443 is correct for Windows, 8343 is for Unix, at least according to the Documentation.

Options for Onedrive integration with SAS Studio/EG

snoopy369 — Fri, 27 Aug 2021 16:26:16 GMT

My organization has started the move to Onedrive/MS 365, and we're trying to figure out the best way to utilize SAS with files stored in Onedrive or in Teams shared storage.

Prior to Onedrive, we did have Sharepoint as a location people could store files, and we didn't get requests to "open them in SAS" because nobody had a share mapped to a local folder (possible to do, of course, using WebDAV, but not common here). However, now that we all have our shares mapped to a local (c:\whatever\Onedrive) folder, we've had requests to be able to store non-sensitive data in these spaces but still be able to access them via SAS.

What are our options here? I have watched and read Chris Hemedinger's article on using the MS Graph API, and that seems like one possibility, but I tend to consider that "high difficulty"; while I certainly can, and will, use this, I suspect most of our users aren't technically savvy enough to implement this, even if I do as much of the work as possible for them - even if we get past the auth token step, the PROC HTTP access to files will probably be a level of difficulty past what they're comfortable with.

Our users are mostly "base SAS" users, who were converted to SAS Studio (90%) or Enterprise Guide (10%) using a 9.4TS1M7 SAS server a year or so ago. They're mostly doing data step/basic Stats procs, importing excel files either as actual data or as metadata to drive processing. Mostly just making SAS datasets, or basic reports (probably an exported dataset or a proc freq/print).

How do others solve this problem? Our ideal solution would be to have a network folder mapped to their Onedrive share and/or synced to Onedrive, but that doesn't seem to be possible. (It might be possible to do it using symlinks, but I don't think that's a good idea.) I also saw that it's possible to map a network drive using a URL, but it seems like that's pretty hack-y also and might not be possible for us, as it requires individual client tokens (we'd have to know what client token to use in the level_env_usermods.bat that would do the mapping, and I don't know how we'd do that without some sort of unsafe storage). We strongly prefer supported, and intended, solutions.

Possibilities I have come up with so far:

Solution	Difficulty	Feature
Graph API	High	Full ability to access files with no manual intervention
Copy files from c:\whatever\onedrive to \\network\share	Low	Entirely manual
SAS Office Add-In	Medium	Can do some things, but we don't use the cooler features, and probably more work to do the basic things

What else is out there? Anything better? Or is this just "how it is" for Onedrive, and people shouldn't use the shared storage for anything they want to use with SAS?