You need a map for SAS with hierarchical sources (XML, JSON). Look at this paper as a guide: https://www.sas.com/content/dam/SAS/support/en/sas-global-forum-proceedings/2018/1734-2018.pdf ... View more

A lot of times, these large, exponential increases in time are due to network latency. The dataset is on a NAS drive or SAN, and you are trying to pull the data to your local machine across VPN. Do as suggested and force the processing to the server or some place local to the dataset. Ping the server where the dataset is located and you will probably find your culprit. 28M recs is not that much but the time indicates a lot of back and forth over a network.  ... View more

That doesn't make a lot of sense.  How are you reading the file back in? It doesn't sound like an issue with the file transfer but the way the file is being read on the dest side. ... View more

Here is a sample multi request with proc http. You need SAS 9.4M4 or higher:   /*=================================================== | USAGE: OPERATIONAL CURL | ----------------------------------------------------- | | curl --location --request POST 'http://localhost:9003/Email/SendEmailAsync?toAddresses=alan.churchill%40savian.net&subject=PostMan' \ | --header 'accept: text/plain' \ | --header 'Content-Type: multipart/form-data' \ | --form 'files=@"/E:/temp/Secured/work/message.html"' \ | --form 'files=@"/E:/temp/Secured/work/QA.xlsx"' *====================================================*/ filename xcl "E:\temp\Secured\HCAHPS\work\Analysis.xlsx"; filename input "E:\temp\Secured\HCAHPS\work\message.html"; filename debug "e:\temp\SasHttp.txt"; proc http url="http://daasserver01:9003/Email/SendEmailAsync" query = ("toAddresses"="alan.churchill@savian.net" "subject"="SAS Email") headerout=debug method="POST" in = multi FORM ( "message.html" = input header="Content-Type: text/html", Analysis.xlsx" = xcl header="Content-Type: application/vnd.openxmlformats-officedocument.spreadsheetml.sheet" ); DEBUG LEVEL=3; run; ... View more

I work with Excel and SAS dataset at every client engagement and almost daily.   I never use SAS to export to Excel. SAS datasets are ODBC and OleDb compliant. Read the data into an Excel technology and do everything in alternate code. You can use C#/VSTO, or C#/EPPlus, C#/Excel interop, C#/GemBox (cost $). I would recommend not using VBA, Python, or Java. Use MS tech to work with MS tech. There are loads of examples out there. You can download mine, for free, here: GitHub - savian-net/ExcelSasDemo       ... View more

SAS Help Center: SAS Procedures by Product ... View more

It depends on what you mean. If you want to know if it works, use SASKiwi's approach. if you want to find it in the project. I just unzip it and then do a regex search for it. I have a web service that unwinds an EG solution to a temp dir, then scans it using regex for all proc/data step boundaries. It kind of depends on what you want. ... View more

You need the ODBC driver install for the db you are using (sql server). If you are using EG via a server, it needs to be installed there, ax others have stated. ... View more

I put that in my first example. Trick is to make sure you get the ODBC connection string exact. You can get that  using the udl trick. the udl trick, if followed all the way to where the string is stored, will provide the entire connection string for you. Otherwise, copy the driver string from the ODBC wizards in Windows. ... View more

If you use that UDL trick I mentioned, on the server, or go into the ODBC drivers and look, you can get the ODBC driver version that is on your system. The UDL trick combined with a far easier way to connect using the ODBC NOPROMPT option will help you find the issue. ... View more

Rather than using the built-in variables, try a direct connection string. This keeps you from having to put the name in place. Also, try the udl file trick (search or just look here: Dan Swearingen's Tech Stuff Blog » Blog Archive » Using a UDL file to generate and test database connection strings (polyweb.com):   libname SQLREF ODBC NOPROMPT="server=Server01.mycompany.com;driver=ODBC Driver 17 for SQL Server;Trusted_Connection=yes;Database=MY_DB;"     STRINGDATES=NO IGNORE_READ_ONLY_COLUMNS=YES SCHEMA=DBO BULKLOAD=YES; proc sql;    SELECT DISTINCT SampleMode   FROM SQLRef.Sales2016   ; quit;   ... View more

Also, Sqlite is not related to Python. It is a nice system because it is a single file database. For SAS data, if I recall correctly, I get an enormous savings in space vs a SAS dataset. It has its issues such as minimal typing (ex. no datetime). ... View more