That's a fast drive. A quick web search shows that this drive needs a "firmware upgrade later than 0B2QJXD7 version, as there is a bug that causes performance issues and premature wear. Use the Samsung Magician software to check the firmware version and update it, if necessary." ... View more

This used to be common with SSDs, where while the SSD does its internal space management, it can slow down significantly.  Read about wear leveling and TRIM to know more. Recent SSDs and operating systems address this well though, so seeing this nowadays is uncommon.  Another reason could be something to do with how the SSD manages its cache. Again, this issue is now much rarer than it used to be.  Note that writing 5 GB in 3 seconds is very fast, so I'd say that the most likely explanation is that the OS caches the data in memory. When the cache is full, actual disk operations take place and that's much slower. 5 GB should take 10 to 15 seconds to read and then the same to write on a SATA SSD, so 30 s seems about right. SAS issues exactly the same code each time, so I doubt the issue is SAS-related.   If you want faster I/O, replace your SATA SSD (if that's what you have) with a M.2 SSD. ... View more

Have you tried using SQL passthru to use the 4-part identifier reference ? Like: proc sql; connect using SQLLIB; select * from connection to SQLLIB ( SELECT * FROM [.\LNKSVR].master.sys.databases ); quit; ... View more

  Do both libname definitions include the database name? What about the ODBC driver definitions?   ... View more

Moved to SAS Viya community ... View more

1. Just a note If you have SAS/Connect, you have this functionality by using MP connect   2. Please check file names are like 'Test.sas' for log file or 'Test.sas' for lst type file. It looks like you have file extensions hidden in Windows. Disable that nonsense.     ... View more

You need to provide an example we can reproduce, and explain where your expectations are not met. ... View more

From the documentation: about Universal Printing: SAS supports TIFF 6.0 for RGBA and CMYK colors. When SAS creates a TIFF image, the image is compressed. If the NOUPRINTCOMPRESSION is set, the size of a TIFF image file that SAS creates is extremely large. Have you tried that syntax and that option? ... View more

It looks like people have been overwhelmed by the wall of text, and have no time to spend deciphering your complex explanations.  Try simplifying what you want to achieve to increase uptake. ... View more

  I never have to wait 5 minutes, but 5 to 15 seconds is common. That's much too slow to load a page   The SAS documentation website is the slowest website I use by a wide margin.   Also, the two screens (user id then password) to login are annoying.   ... View more

Performance is fickle and depends on the data and how you use it. ALTER creates a new data set, and isn't usually faster than CREATE, but it depends.. SQL does things in a different way, and can also exhibit slowness compared to data steps. SAS has really tuned these over the years. Consider this: data TEST1 TEST2 TEST3; do I=1 to 5e7; output; end; run; proc sql; %*** 4.1 seconds ***; alter table TEST1 add J int; quit; proc sql; %*** 4.7 seconds ***; create table TEST2 as select *, . as J from TEST2 ; quit; data TEST3; %*** 2.7 seconds ***; set TEST3; retain J .; run; A data step is fastest here, CREATE takes 70% more time. ALTER is in the middle. I ran that code on SAS ODA btw.   Bottom line: You'll build a feel for what methods work best for different tasks, but when performance becomes important, nothing replaces bench-marking your processes with your data on your hardware.   ... View more

Well, if you are using 9.4, then you need another application to create the parquet files for you.  You can use Hive as you seem to be doing now, or use a Python package such as Panda that you can call from SAS, for example. Then you transfer the files to S3. There might be other methods (such as AWS Glue), but I'm not familiar with them.   Note that option BULKLOAD allows loading multiple rows of data as one unit, to insert or append them to a DBMS table. A S3 bucket is not a DBMS. You are not shifting rows, you are shifting files, so this option has nothing to do with S3. Caveat: I never used S3, so my understanding of its capabilities might be limited. ... View more

You need to provide more details. Are you using 9.4 or Viya? Do the parquet files already exist? SAS 9.4 cannot create parquet files.  Viya can, see here, and see here for an S3 upload example. ... View more

This question has been asked before. Have you seen this? ... View more

I don't see where you use a macro variable. In any case, the message ERROR: CLI describe error: [IBM][CLI Driver][DB2/AIX64] SQL0101N The statement was not processed because a limit such as a memory limit, an SQL limit, or a database limit was reached. SQLSTATE=54001 indicates that the statement is too long or too complex.  This is a DB2 limitation: Your SQL clause is too long or complex and DB2 rejects it. See https://www.ibm.com/docs/en/db2-for-zos/12?topic=codes-101 Talk to the DB2 DBA for recommendations. As stated, the simplest work-around is to join to an uploaded table. Another solution is to send a query smaller than the maximum allowed, and apply the rest of the tests in a subsequent phase.         ... View more