Here is an outline a macro .  Basically you want to put almost all of your code into a macro, lets call it MorisitaIndx.  Added code to make it generic is in bold. %Macro MorisitaIndx (forest) ; data usedata ;      Set data (where = (forestid = &forest)) ;  /*  data will be read outside macro */ run ; *start Morosita's index computations; proc iml; /** Count the number of points in each cell of a 2D grid. .... */ .... Proc IML .... use usedata ; *bring external data into IML environment (SAS data but not data curently inside IML procedure); ... data morisita;   Retain forestid &forest ; /* use if your code does not keep the forestid variable */   set morisita;   MInd=sumc*_freq_/(points*(points-1));   pvalue=1-probchi((mind*(points-1)+_freq_-points)/(_freq_-1),_freq_-1); run; proc print data=morisita; run; Proc Append      Base = AllMorisita      data = morisita      ; run ; %Mend ; data data;   infile "B:\sas\data.csv" delimiter=',' firstobs=2; * read the csv data with coordinates of the trees and multiple forests;   input forestid $ x y diameter;      /* delete this next line */   * if forestid ne 1 then delete; * select only one forest; run; /* List the forests */ Proc SQL ;      create table forests as      select distinct forestid      from data Quit ; /* Run through the macro for each forest */ Data _Null_ ;      Set forests ;      ExecStr = cats ('%MorisitaIndx (' , forest, ') ;' ) ;      Call Execute (ExecStr) ; Run ; Warning - untested code - you might have to adjust the data _null_ step Richard ... View more

I don't understand why the first record in your 'want' table for id 11111111 picks up date and cat data for id 33333333 a couple of rows down.  Assuming these are mistakes it seems you want to suppress key information from the first few columns where the data is repeated in succeeding rows.  This would be a bad idea for storing data for future analysis but might make sense if what you want to do next is put the 'want' data into a report. Assuming your data is sorted by maturity amt the following data step should achieve the result: options missing = ' ' ;  /* use blank instead of . for displaying missing numerical data */ /* check syntax */ data want ;      set have ;      by maturity amt ;      if not first. maturity then maturity = . ;      if not first.amt then amt = . ; run ; Note that data is no longer sorted because values of key variables have been suppressed. Richard ... View more

You may just need to change the format in Oracle.  As far as recall, Oracle stores all dates as datetime values. Richard ... View more

A datastep merge might be simpler, assuming the data is ordered as shown: Data want ;      Merge    have (where = (Rec1 is not Null)                            drop = Rec2 orderRec2                           )                   have (where = (Rec2 is not Null)                                                    drop = Rec1 orderRec1                           )                   ;      By     ID ; Run ; Richard ... View more

data want ;      set have ;           target1 = (a1=5) * (a2=9) ; etc ; run ; Richard ... View more

At a guess it looks like your col1 represents a percentage (not just amounts with a percentage format).  Hence to find the matches you need to multiply col2 by 100 (or divide col1 by 100 - but for rounding reasons multiply col1). You may find that there are still rounding issues in your data so you may also need to round before comparing or accept a difference of <.01 as denoting equal. In its simplest form      data equal (where  (col1 = 100 * col2))           unequal (where (col1 <> 100 * col2))           ;           Set have ;      run ; I would suggest a more robust version to test the differences      data equal (where  (test = 0))           unequal (where (test > 0))           ;           Set have ;           test = ABS (col1 - 100 * col2) ;      run ; Then test the unequal differences ;      Proc univariate           data = unequal ;           var test ;      run ; This should give you enough information to set a minimum value for test for which you will accept inequality. (ie change the condition test = 0 to test > .05 etc) Richard ... View more

DDE is specific to the Windows environment. You will have to use proc export to export your list data Be aware that there is a bug in SAS 9.1.3 - fixed in 9.2 (20325 - Using the EXPORT procedure with DBMS=XLS in 9.1.3 SAS/ACCESS® Interface to PC Files under UNIX to create a Microsoft Excel file might cause problems with the Excel file) In the past (in Windows) I have had to fiddle with tab names to make export work - SAS often created a new tab instead of updating the existing one.  Best idea was to let SAS create the list tab then open the workbook in MS Excel to create the charts for the first time. Another approach, if you have a Windows PC with a full Base SAS installation and SAS connect in both environments is to use the Windows PC as the server (rsubmit from UNIX, opposite to the usual config).  You should be able to rsubmit code using DDE or a windows libname to update your workbook (which has to be open in Windows for DDE to work) and do the update that way, if necessary then uploading the finished workbook to your UNIX server. Richard ... View more

If your data ranges are always the same, even if the list data is refreshed, just create the charts in Excel.  They will update when the list data is refreshed. If there are parameters which need to control how the charts are displayed (eg titles) you may be able to define and export those to a hidden tab, to be accessed in the chart. Richard ... View more