I think you need a macro variable which contains the minimum of the data. From there you can determine how to set MIN= ... View more

If change the code as below, y2axis min=0; The plot shows up with y2axis=0. But the min of y2axis could be any value.  Does this mean the min of y2axis could be negative? What if the min of y2axis is 75? Do you still want 0 to appear on the axis?   It helps if you explain clearly what you DO want, rather than stating it doesn't work in some situations. We need rules about what should appear in every situation, not examples of a few things that are wrong. Rules, not examples. ... View more

@Akshaya_1397 wrote:   3.When i check using =value() it displays mm/dd/yy format. What does this mean? I don't understand. Please provide more details.   4.when i exported to sas it shows as mm/dd/yy format as numeric 8. How did you import this into SAS? What steps? What code?   What happens if you change the format in SAS to a datetime format, like DATETIME16. ?? ... View more

COMPRESS option does save space, but it requires more CPU time.   Regarding setting the length to 3, if all the values of every one of the 200 variables wind up in the new dataset value column, you can't have length 3, you will most likely need length 8 unless ALL original variables fit in that length.   IMHO, estimating the new size of the dataset is a waste of time, as I said, the new data set is almost useless in that form. ... View more

Lets say that I consider perform transpose so there will be Long structure with 3 fields: Custt_ID,var_name, value   I don't know the answer to your question about file size, but this arrangement of the data after the transpose will be a nightmare to use in further programming. Furthermore, when we talk about long data sets in SAS, what you describe isn't what we mean (although in non-SAS usage of the word "long" it certainly is long). ... View more

Use the %QLIST macro https://github.com/sasutils/macros/blob/master/qlist.sas ... View more

compress(var2,'', 'A')   The COMPRESS function produces a character string, in this case a character string of the letters in VAR2. For the first row of your data set you get the character string '01'. It is not a number, even if it looks like a number to you, so SAS does not know how to evaluate if the character string '01' is greater than &range_1, which is a number. SAS cannot compare characters strings to numbers. So to fix this, you want to convert the character string from var2 into an actual number. You do this by   input(compress(var2,'', 'A'),3.)   which turns the character string into the actual number 1. Then it can be compared to the number in &range_1. This will handle numbers up to 3 digits wide (i.e <= 999).   My macro variables are defined as: %let range_1 = 13; %let range_2 = 24; so they appear to be numeric so I'm not sure where my error might be?   The values of macro variables are always text, no matter that they look like numbers. The only way they become numbers is if you use them in a way that SAS recognizes them as numbers. In this case, when SAS evaluates the code, it replaces &range_1 with 13, and so then the code actually becomes    compress(var2,'', 'A') > 13   and then SAS recognizes 13 as a number from the context. Do you know why in the above line that 13 is treated as a number and not a character string?   @Jeff_DOC you're not new here. Please be a good citizen of the community and help us just as we are trying to help you. When you get errors in the log, then SHOW US THE LOG. We need to see the entire log for the DATA step (or PROC) that has the error from now on, without having ask. ... View more

Is this variable that goes from 1 to 96 numeric or character, according to PROC CONTENTS?   If it is numeric, what is the format according to PROC CONTENTS? ... View more

Is variable X numeric or character, according to PROC CONTENTS? The results you are getting make me think it is character. If it is character, then you will need to convert it to an actual numeric date/time value, assuming that you want it to be a date/time value (is that what you want? you tell us what you don't want, but you haven't really been specific about what you do want). ... View more

And what analysis will you be doing? When I look at the examples in PROC PANEL they are all long data sets, not wide data sets. Most SAS data analysis PROCs are designed to use long data sets. ... View more

So @Ksharp has already given you the answer, however, I would like to point out an important principle. Best practice is to handle dates using date functions and date formats and date informats, if at all possible. And to do this, you need to know if the value of End_Date (which you claim is '2014-10-04') is really character, or really numeric. If, as Ksharp guessed, the variable End_Date is character (is it?) then you need to use an informat (which in Ksharp's code is yymmdd10.). If End_Date was numeric, then you could use a format to convert it to the desired appearance. ... View more

IMHO The best thing to do with this data set is to leave it as a long data set, and perform any analyses on the long data set. But you probably knew that already. The less processing you have to do on 17 million records, the better off you will be.   Please explain why you want it to be a wide data set. ... View more

Please explain further. How can we calculate annual change if there is no variable in your data set that indicates year?   Please explain further. You have weighted data, what are the weights, and which variable(s) are weighted, and how/why would we use the weights? ... View more

These are confidence intervals with respect to what randomness? No population or sampling from that population has been discussed by the original poster.   These numbers that @SASUser67 talks about could be non-sample from a population, they could be the entire population, in which case there is no randomness (and no confidence intervals)   I think @SASUser67 needs to explain a whole lot more before we can say that a Poisson distribution applies here. ... View more