It would be cool if a proc existed that output macro documentation of a compiled macro. Kind of like Proc Catalog but in addition to the current fields it provides – it included a more robust description (more than 40 characters) and included the list of keyword or positional parameters. Maybe even info like if the parameter is required, optional, default value if defaulted, example value, and an example macro call. Similar to Proc Contents where you can turn on/off the level of details requested - you can get a high-level list of tables & their values, and a detailed page of each individual table. It would be helpful if something similar existed for compiled macros. ... View more

A long time ago, we all learned that 2-digit years are bad.   SAS itself is Y2K compliant, but if user-written code (or data) has two-digit years (or one-digit years!), SAS is left to impute the century by referring to the setting of the YEARCUTOFF option.  Below illustrates how YEARCUTOFF can be used to impute century (and even decade):   1 options yearcutoff=1800 ; 2 data _null_ ; 3 x1='7JUN2018'd ; 4 x2='7JUN18'd ; 5 x3='7JUN8'd ; 6 x4='7JUN8fooooo'd ; 7 put x1= x2= x3= x4= ; 8 format x1 x2 x3 x4 date9. ; 9 run; x1=07JUN2018 x2=07JUN1818 x3=07JUN1808 x4=07JUN1808 I think it would be helpful to have an option like YEARCUTOFF=ERROR or NOYEARCUTOFF.  This would say "If SAS encounters a year with less than four digits, generate an error message, rather than using YEARCUTOFF to impute the century."   While it can sometimes be helpful to write code that is tolerant to two-digit years, it would be useful to be able turn on stricter error handling which requires all years to be four digits.   Related discussion thread: https://communities.sas.com/t5/SAS-Programming/YEARCUTOFF-ERROR/m-p/468377/highlight/true#M119620 ... View more

It's very helpful to name your graphic images in a consistent way when creating items in SAS/GRAPH with a name statement.  Currently we are limited to 8 characters.  It would be great to go to 32 characters as for variable and data set names.  I'm old enough to remember when we were limited to 5 characters for variable names on some antiquated systems (i.e. UNIVAC) so I'm pretty good at trimming down names but I'd rather not have to! ... View more

It would be useful to have an informat like MDYAMPMw.d but for us non-US users. ... View more

At the moment, quote() can only add enclose a string in ' or ". It would be useful if any character could be added. And if the quoting character is one of < > [ ] { } ( ), this function would also know to enclose the string in balanced brackets.   A 't' third parameter to trim the quoted string would also be useful. ... View more

The format $QUOTE will enclose the value in double quote characters and properly handles doubling up any embedded double quote characters..  It would be useful to have a format that would generated a quoted string using single quote characters.  String literals quoted using single quotes will suppress any embedded macro triggers and can be used in pass-through SQL code.   Perhaps they could add a letter to the end of the format name like they have done with the date formats. ... View more

All tables are indexed, so this takes no time.   data ACCOUNT_444513219790; set XODSM.ACCOUNT_201101-XODSM.ACCOUNT_201112 XODSM.ACCOUNT_201201-XODSM.ACCOUNT_201212 XODSM.ACCOUNT_201301-XODSM.ACCOUNT_201312 XODSM.ACCOUNT_201401-XODSM.ACCOUNT_201412 XODSM.ACCOUNT_201501-XODSM.ACCOUNT_201512 XODSM.ACCOUNT_201601-XODSM.ACCOUNT_201606; where IP_LOCAL_ID='444513219790'; run;   If I use a view, it takes minutes since the where clause is processed in the second data step.   data _V_ACCOUNTS/view=_V_ACCOUNTS; set XODSM.ACCOUNT_201101-XODSM.ACCOUNT_201112 XODSM.ACCOUNT_201201-XODSM.ACCOUNT_201212 XODSM.ACCOUNT_201301-XODSM.ACCOUNT_201312 XODSM.ACCOUNT_201401-XODSM.ACCOUNT_201412 XODSM.ACCOUNT_201501-XODSM.ACCOUNT_201512 XODSM.ACCOUNT_201601-XODSM.ACCOUNT_201606; run; data ACCOUNT_444513219790; set _V_ACCOUNTS; where IP_LOCAL_ID='444513219790'; run; I would be better if the where clause was processed by the view.   Suggestion: Any where clause on a view should be passed on for processing by the view whenever possible.   An example of a case when this is not possible is when data sets include the where=option. Another example is when the point= or key= options are used in the view. ... View more

  data T(sortedby=X); X=1; output; X=2; output; run;   will not set the validated flag in the data set's metadata as seen here:     One must recreate the table using for example proc sort to set this flag. This is hugely wasteful.   The following option would ask SAS to check the sort order as the data is written, in order to avoid this waste:   data T(sortedby=X validatesort); X=1; output; X=2; output; run;                   ... View more

The datetime output is identical to that of the E8601DTw.d format except for the T separator.    "yyyy-mm-dd hh:mm:ss.ffffff"  instead of  "yyyy-mm-ddThh:mm:ss.ffffff"   A format generating the same output for date (and time?) variables would be useful too as some tools like Impala can only use timestamps, not dates.   I know this can be done using a picture format. Doing so makes the code a lot less portable though.   ... View more

Just a simple function to replace   TMP=A; A  =B; B  =TMP;   with shorter and more legible   call swap(A,B);   and RC=swap(A,B); for macro use.     ... View more

It's frustrating when you write a %macro function and/or a %do loop, and the DATA and PROC steps inside are not colored according to the usual syntax highlighting.   This was suggested as a feature back in version 9.1, but it still hasn't been fixed. ... View more

This would allow code like   %if &i=1  ## comment 1   | &i=2  ## comment 2   | &i=3  ## comment 3 %then %let i=4;   The only method for these comments atm is to use /* */ comments, which I avoid as I reserve /* */  to comment out large blocks of text on an ad-hoc basis (debugging, rerunning parts of a program).   And of course blocks of text cannot be commented out if they contain /* */ comments themselves. If they do, the only way to comment them out if to enclose them in dummy macros, which can cause other issues.   [Edit: Replaced // with ## in my example, as // is commonly used to skip lines in put statements. I don't mind what characters are used.]   ... View more

The ability to have multiple cursors would be a huge improvment. Many other editors have this option (specifically Rstudio and Sublime) and adding it to SAS would bring a lot of people back to SAS as their main editor. The uses for this feature are nearly infinite! ... View more

On hierarchical filesystems, a directory is just a file (albeit of a specific type), but nonetheless a file that has the same attributes any other file has: permissions, modification or other date(s), a size. The directory-specific functions DOPEN() et al are suited to mostly process the contents of the directory, but they miss out on the file-specific attributes of the directory file itself. I therefore propose to enable the FOPEN() and other functions to work with a directory file for the purely FILE aspects of it. ... View more

Currently, just a few options (they don't all appear when running proc options group=  errorhandling; ) allow to decide how message are displayed in the LOG (I am not too sure what the last 2 do...), like:   MERROR             Issues a warning message for an unresolved macro reference. SERROR             Issues a warning message when a macro variable reference does not match a macro variable. VALIDFMTNAME Specifies the maximum size that user-created formats and informat names can be before an error or warning is issued. BYERR              SAS issues an error message and stops processing if the SORT procedure attempts to sort a _NULL_ data set. DKRICOND    Specifies the error level to report when a variable is missing from an input data set during the processing of a DROP=, KEEP=, or RENAME= data set option. DKROCOND Specifies the error level to report when a variable is missing from an output data set during the processing of a DROP=, KEEP=, or RENAME= data set option. DSNFERR Issues an error message and stops processing when a SAS data set cannot be found. MERGENOBY Specifies the type of message that is issued when MERGE processing occurs without an associated BY statement. VARINITCHK Specifies the type of message to write to the SAS log when a variable is not initialized. VARLENCHK Specifies the type of message to write to the SAS log when the length of the variable that is being read is longer than the length that is defined for the variable.   DS2SCOND Specifies the type of message that PROC DS2 generates. RSASIOTRANSERROR   Displays a transcoding error when illegal values are read from a remote application. VNFERR             SAS issues an error message when a BY variable exists in one data set but not another when the other data set is _NULL_.   We need more ways to customise how messages are displayed in the LOG when some events are flagged.   Some of the messages we don't consider worthy of a warning (some flag something that can't even be fixed!) include:   WARNING: The intervals on the axis labeled XXX are not evenly spaced WARNING: The enhanced date axis procedure has failed WARNING: Compression was disabled for data set WARNING: FOOTNOTE... WARNING: TITLE...   They give us a hard time for monitoring automated production jobs.   On the contrary notes like these should be show-stoppers:   NOTE: MERGE statement has more than one data set NOTE: Library  XXX does not exist. NOTE: SAS went to a new line NOTE: DATA STEP stopped due...   It would be nice to be able to raise valid alarms and avoid false alarms by setting the type of message for the various messages that SAS produces. ... View more

Hello,   Mirroring the WORKPERMS system option, valid for SAS UNIX/LINUX    https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/hostunx/p1bd28x2b4hcwin1b68clthsekf8.htm   A new system option - let's call it USERPERMS , would be very useful to harden within SAS the security of the SASUSER directory path the first time it is created.   syntax :   USERPERMS < POSIX security profile 700, 775, 750 etc.>   ex. -USERPERMS 750  creates the directory with the following permssion profile drwxr-x---. 2 userid group 4096 Apr 21 14:34 sasuser.v94   Valid in configuration files   The workarounds exist, of course, but they entail the execution of sensitive system commands like chmod in an automated script, which is not a good practice in my experience. Permission assignments should never be run automatically, thats a SPOF large enough to make way for uninvited accesses when the automation itself malfunctions . Instead, use umask initial value setting like ... WORKPERMS , or soon USERPERMS ?    Comments are welcome,   Best Regards,   Ronan   PS. Sorry if this request duplicates another , I was unable to find a previous one. ... View more

https://communities.sas.com/t5/SAS-Data-Management/How-do-I-get-regular-expressions-functions-in-PROC-FEDSQL/m-p/472453#M14809   PROC DS2; package regexp /overwrite=yes; method match(varchar(1024) pattern, varchar(1024) string) returns varchar(1024); return prxmatch(pattern, string); end; endpackage; run; QUIT; PROC FEDSQL; CREATE TABLE foo AS SELECT regexp.match('/[0-9]+/', id) AS test1, regexp.match('/[A-Z]+/', id) AS test FROM someDatabase.someSchema.someTable; QUIT;   With an error message of:   ERROR: Access Violation occurred during PREPARE! Access Violation? is the error that needs to be better explained. ... View more

We have many datasets with formats attached that we don't normally use. So we load them with NOFMTERR in effect.   However, this clutters up the log file with lots of notes about formats that could not be found. Surely these are not needed? Alternatively, if you don't want to break backwards consistency introduce another option like NOFMTERRNOTE which suppresses these, or add a dataset option to not load the formats for a dataset. ... View more

I'd like to see a simpler way of creating and using hash tables...something consistent with the syntax and structure of either data statements or proc sql structure.   My suggestion...something lie this: libname foobar memory;    (kind of like memlib, but cross-platform)   and then it could be generated like this: create foobar as select id,desc from sometable key=id; -or- data foobar; set sometable; keep id,desc; key=id; run;   from that point on, foobar could be used just like any other dataset: proc sql; select store,storetable.id,foobar.desc from storetable left join foobar on storetable.id=foobar.id;   That seems much easier to use and teach than hash and it's yet-another-syntax.   ... View more