Maybe you should look at FORECAST and ARIMA procedures (or some forecasting tasks in Enterprise Guide) : those procedures can model your time series and predict the future with a growing confidence interval as times goes by... Unfortunately, I'm not keen on time series for myself. ... View more

Thank you Cynthia. ... View more

If you are using SAS on Windows, you should be able to use the Enhanced Editor, and therefore to record a keybord macro : action = type /********************** and associated key = F11. If on a different OS, the ~ trick Cynthia just explained should work alright ; the keys are save on a profile basis, so you should have the F11 and F12 re-definitions saved when you close your SAS session. ... View more

Hello. In PROC GLM (as well as other modelling procedures), the * sign does not mean "multiply" (as I think you thought) but "cross". X2*X3 is cross-effect for X2 and X3, which means groups defined by values of BOTH X2 and X3, and not groups for X2 on one hand, and groups for X3 on another hand. Imagine you have a study of respiratory capacity (lung volume or something like that) recorded for patients, some of whom are smokers, other non-smokers ; some are men, other women. When you just test the SEX effect, you will know if lung volume is different for men and women (it usually is). When you just test the SMOKE_HABIT effect, you will know if smoking decreases the lung capacity, regardless of sex. What you test when considering the SEX*SMOKE_HABIT is if among women and men taken separately, smoking alters the respiratory capacity. I hope it is clear enough for you. ... View more

I'm afraid there would not be any generic answer to that kind of problem. Yet I can suggest some hints to help : 1) DATA ... ; INFILE ... ; INPUT ; After that, a temporary character variable named _INFILE_ is available in the PDV. It consists in the whole line of data. 2) you can then try to "slice" it down into variables using REVERSE, SCAN, PRXMATCH, INDEXW functions. 3) you can test "patterns" in data (such as a number, with or without commas), with PRXMATCH and/or PUT functions. That would be a painful job, but if you can't get your data source add a real delimiter into your file, I'm afraid you'll have to go through this. ... View more

I suggest you replace your OUTEST= option with an ODS OUTPUT statement prior to PROC LOGISTIC, this way : ODS OUTPUT parameterEstimates = work.myCoefficients ; PROC LOGISTIC etc... You will have in the MyCoefficients dataset both parameter values and associated Wald tests statistics & p-values. ... View more

Hello. There is no real annotate feature outside SAS/GRAPH, so you can't really annotate a proc plot. Either use GPLOT or consider the following syntax : PROC PLOT DATA = ... ; PLOT Xvar * Yvar = "#" $ labelVar ; RUN ; QUIT ; THe # symbol will be used to mark points, and the values of labelVar will be written next to the points. For "square" graphics with both axis of the same length, try to change the LS and PS option to control the Output page size. You can shrink or expand axes that way (you can do the same with SAS/GRAPH using the VSIZE and HSIZE GOPTIONS). But it's not a very satisfying way to do things, just a workaround. Maybe you can check also the ODS GRAPHICS templates : maybe there is a "square plot" template that you could use in a Data step (I've seen SUGI papers talking about such things). Hope it helps. Olivier ... View more

What you are looking to is called "classification table". You should try the CTABLE option (in the MODEL statement options) ; this is not really cross-tabulated, but contains all information you need. You can also create predicted observations with the OUTPUT statement, then create a format with PROC FORMAT to transform continuous predicted probabilities into binary predictions (e.g. 0-0.5="0" 0.5<-1="1") and use both table and format in a FREQ procedure. ... View more

I am not sure to understand your problem properly : - does your GBARLINE proc only produce ONE and only one bar (then the previous answer should fix ; you can either use the DISCRETE option in the BAR statement to get one bar per value) ; - or do you want two SERIES of bars on the same graph, in an overlayed kind of graph ? ... View more

Hi, there is a LARGEST(k, variable_list) beginning with SAS9 ; k equals 1 if you want the maximum, k=2 will give you the second greatest value, and so on. For your problem of recording the variables names according to their order, is the following example of any help ? DATA work.test ; INPUT x1-x5 ; CARDS ; 45 454 484 21 2 2121 5156 5645 451 5451 5151 2121 21 313 26 55 64 65 1 231 ; RUN ; DATA work.whos_who (DROP = nth: j) ; SET work.test ; ARRAY myXs _NUMERIC_ ; ARRAY theirNames $ name1-name5 ; DO nth = 1 TO DIM(myXs) ; nth_biggest = LARGEST(nth, OF x1-x5) ; DO j = 1 TO DIM(myXs) ; IF nth_biggest = myXs(j) THEN theirNames(nth) = VNAME(myXs(j)) ; END ; END ; RUN ; ... View more

You can define your library with a -SET myLibNam "path" in the config file of your SAS server, instead of a LIBNAME statement in an AUTOEXEC or a definition with the Enterprise Guide Administrator / Management Console. The library should actually exist, but remain invisible to users. However, they will be able to use its contents through programs, so maybe you should keep passwords on the datasets too. ... View more

You can follow these steps : 1) create a query on your dataset 2) in the SELECT/SORT tab, keep all the columns you need at the end 3) in this tab, choose a value in the SIGMA drop-down list (just right to the check boxes) for the variable of which you want to use frequencies 4) drag and drop one more time this particular variable in the SELECT/SORT TAB, and give it a value in the SIGMA list 5) rename this last column COUNT 6) go to the GROUP FILTERS tab, and build a condition using the newly created COUNT variable (should be at the very bottom of the list on the left). I hope it solves your problem. ... View more

You should consider using this little program... (In a Code window.) %LET dsIn = ... ; /* your existing DATETIME dataset*/ %LET dsOut = ... ; /* your converted dataset ; you can use the same name twice, and you will have your existing dataset replaced */ /*************************************************/ /* DO NOT EDIT ANYTHING AFTER THIS LINE */ /*************************************************/ PROC CONTENTS DATA = &dsIn (KEEP = _numeric_) OUT = work.dictionary NOPRINT ; RUN ; PROC SQL NOPRINT ; SELECT name INTO : varToConvert SEPARATED BY " " FROM work.dictionary ; QUIT ; DATA &dsOut ; SET &dsIn ; ARRAY datetimes &varToConvert ; DO OVER dateTimes ; dateTimes = DATEPART(dateTimes) ; END ; FORMAT &varToConvert DDMMYY10. ; RUN ; ... View more

You can try : 1) creating a character variable with the formula PUT(myDate, MMYYS7.) and your new variable will only contain month and year 2) create a new date variable with the formula INTNX("MONTH",myDate,0) and you new variable will be located on the 1st of the month ... View more