About ballardw

ballardw · ‎02-25-2026

Best practice when code generates errors is to include the LOG with the code and error, warning or other messages of concern. Copy the text from the log, on the forum open a text box window using the </> icon above the main message window and paste the code and messages. The text box is important because the forum software will refomat text pasted into the main message window resulting in some of the diagnostic information that SAS often supplies. As a minimum your code should also include a TYPE variable to tell SAS whether the valus in the START variable are numeric or character and display or input. A type of N (character value of the variable type) means Start is supposed to be numeric and is a display format, Type of C means start is Character and is a display format, Type of I (upper case i) is informat for numeric values, type of J is informat for character values, Type P means a PICTURE format. If your data contains "A1" and "a1" for the variable and want them both to display the same label text then you want both values in the START. Because you IMPORT the data from Excel you may have issues of same value, i.e. "A1" with multiple values of the label cause by capitalization differences or white space in values (different numbers of spaces can be hard to see) or leading spaces. SAS doesn't like that for formats or informats and will show errors of duplicated ranges. Suggestion: Write a small, as in two or there value/invalue statements as normal proc format code and use the CNTLOUT option to write a data set of those. That will show values for the variables needed in the data set. You may want to consider adding and Other = error option to the above example trial code if creating an INFORMAT. The Other option, which would go into a variable named HLO, coupled with the error option means that use of the format with unknown or unexpected values tells you that such a value was encountered in the log with an invalid data message.

ballardw · ‎02-24-2026

You should be able to use Proc Template to get the code for the define style and then recompile it with your current system. I ran into some issues in the past with a custom Style that modified a SAS supplied style and when SAS was upgraded the definition of the same named style changed so that my custom style did not use the SAS style correctly. This code will show the source code for the style Minimal in the log: proc template; source Styles.minimal; run; You could copy that into the editor and recompile the style. Note that I see this in my install's version of Minimal: style Document / doctype = "<!DOCTYPE HTML PUBLIC ""-//W3C//DTD HTML 3.2 Final//EN"">" contenttype = "text/html" protectspecialchars = auto; Which implies that style is intended to create HTML 3 documents which may cause an issue with the Tagset Excelxp. Creating a file with an XLS extension but applied to a different format of file is one of the things Microsoft has made unrelieable as well. The warning messages you show do not seem to apply to my version of Minimal so you may have to chase more things down to use that style and the tagset together. The putvars and newline sound more like table output requests than a general style and may indicate an interaction with the Proc Print default table settings and the minimal style as well.

ballardw · ‎02-20-2026

You will find that having commas in macro variables is often a problem because that is the separator between values in the macro processor and when such macro variables are used in data step processing you will get similar problems except for the functions that lists of values that are separated by commas. Additionally having quote characters as part of the value often complicates things as most functions will expect one value within a set of quotes. You encounter an example with your use of Scan ("&mynames",_n_,',') where the data step will see Scan (""AAA","BBB","CCC","V123"",_n_,','). So the first pair of "" is the empty string to search with everything else problematic as AAA" will not be a valid second parameter because of no comma. Why was your Myname macro variable defined that way? If only to be parsed by this data step then other definitions will likely work better.

ballardw · ‎02-18-2026

@SASdevAnneMarie wrote: Hello Ballardw, "With" spaces. Unfortunately, I can not modify my message. You should be able to edit any of the posts you have made. There should be a block of three lines next to the subject of the post. Roll your cursor over the block of lines and and "edit message" or "edit reply" should appear. Click and the edit option opens the message or reply.

ballardw · ‎02-17-2026

@SASdevAnneMarie wrote: I don't have the missing values in variables. How about all 0 values for Nmbr_de_lgm for one or more departement levels? The CV function has some issues with that I believe.

ballardw · ‎02-17-2026

Beware of _numeric_ with data sets containing date, time or datetime values as many statistics won't be very useful for them.

ballardw · ‎02-17-2026

@SASdevAnneMarie wrote: Thank you, My code is: proc means data=bdd vardef=N mean std var cv median max min q1 q3 maxdec=5; output out=sumstat; run; I do not have cv, q1 ,q3 and median in table sumstat. Once upon a time in SAS there were two procedures Means and Summary. Means sent results to the default output device such as screen or printer. Proc summary sent output to data sets. And the two were separate. At some point the procedures were merged in proc means somewhat. So the default statistics on the procedure statement remain as the statistics sent to the default output screen but those do not affect the default behavior of the output statement which was brought in from proc summary. And don't forget that there are options now that will make output data sets with the ODS OUTPUT statement that will appear in a different form than the Output statement and can have a different structure using the STACKODSOUTPUT option.

ballardw · ‎02-17-2026

@SASdevAnneMarie wrote: Hello Experts, I'm wondering if there is a numeric format without spaces. I apply the format comma10.2 and, for example, my result is 673,380.10, I would like to have 673 380.10. Thank you! I have to say that your comment about "without spaces" confuses me because you say that you "like to have 673 380.10" which clearly has a space between the threes.

ballardw · ‎02-04-2026

A tool that is good to learn for use in debugging macro related code, your %Do loops are macro code, is to turn on one or more options of Mprint, Symbolgen or Mlogic. Add something like options MPRINT; before the code in question when executed. Use Options nomprint; to turn off. You should see that the resolved code in the LOG using the Mprint that you get a separate instruction code line of "run;" because your current' code does not have that quoted string as part of the argument for the PUT instruction. The current "If type= 2" and related code ends the PUT. So you need another PUT before the "run;". Hint: It is usually a good idea to include the entire code for a data step or procedure that is having issues and better to include a small data set in the form of working data step code that could be used to test the code. Since your code doesn't define where the values of Type, X or Y variables come from I have to assume there is another data set involved and more code using that data set prior to what you have shown.

ballardw · ‎02-02-2026

If you are working in an environment with an IT staff you might contact them to see if they made any recent changes to security software or possibly any local or group machine settings. If your organization doesn't have a lot of SAS users the IT staff may have done something that interferes with SAS.

ballardw · ‎01-30-2026

For what very little it may be worth, I tend to create output data sets for Proc Compare and then write a custom report for anything except nearly trivial Proc Compare output because the procedure is quite often not very nice. With a data set then I can use all of the options provided by Proc Print or Report to make sure the appearance is as desired.

ballardw · ‎01-25-2026

Something you might consider would be sharing an example of your code and/or describe what you program is doing in some detail. Often people less familiar with SAS will try to force an approach to some tasks that better done using other procedures in SAS. Also "loops" and "files" in the same phrase could mean that you are doing a lot of disk IO that may not exactly be needed with a different approach.

ballardw · ‎01-20-2026

If you like the appearance that EG gives you then the steps are 1) determine the current ODS STYLE in effect for your session (might be HTMLBlue) 2) on the code that defines the PDF file add the STYLE= option to use the same style. Or if using the EG interface to set the PDF file options there should be a STYLE option somewhere.

ballardw · ‎01-19-2026

Best would be to show an example of the expected result for your "transpose" of the example data. Note that "doesn't do the job" does not provide any detail about what was missing (or extra) or what is actually wrong with the resulting data set. A likely incomplete guess on my part as to part of the objection about the output data set would be the repeated values of the last TypeTransfusion but I am sure that is not the only bit.

ballardw · ‎01-19-2026

ATTRID option specifies the value of the ID variable in a discrete attribute data set. It basically indicates which set ov values to use for a given group variable as the data set could contain the multiple sets of descriptions for the same VALUE Since your code does not show the use of a discrete attribute map data set, which would be indicated by the use of the DATTRMAP option on the Proc Sgplot statement, you question doesn't make much sense as phrased. The default legend would have the values of the Group variable. The color or other options specified for each value would be in the DATTRMAP data set where the ID variable has the value "group2". I would strongly suggest looking in the online help "Using Discrete Attribute Map Sets".

Online Status	Offline
Date Last Visited	yesterday