Hi,   If you wanted to make this a little more automated you could use the following:   data table_1 table_2 table_3 table_4 table_5 table_6; x=1; run; %let keep_table = TABLE_3; proc sql noprint; select memname into :drop_tables separated by " " from sashelp.vcolumn where libname = "WORK" and memname like "TABLE_%" and memname ^= "&keep_table"; quit; proc datasets; delete &drop_tables; quit; This way it doesn't matter how many datasets you create, you only need to specify the single table you want to keep (and if you have a rule for which table you want to keep you can use this to define the "keep_table" macro variable).   ... View more

Hi hx,   I have a solution that fits the data you have presented, however it is dependent on you wanting to group your data based on a match of the first word in the "words" variable (you could edit the wordmtch variable to include the first two words, or write a macro to change this for you, if you also require that) and I'm sure there is a better way to go about doing this, but maybe it will help you.   data have; input words $ 1-28 c1 c2 c3 c4; wordmtch = scan(words,1); cards; ally 21 32 45 65 ally traitor 56 76 45 67 barriar overcome 87 98 21 76 barriar overcome instruction 32 42 12 32 cat 67 36 43 85 cat pat 18 92 18 27 cat pa 82 54 21 34 ; run; data want; set have; by wordmtch; retain c1tot c2tot c3tot c4tot; if first.wordmtch then do c1tot = c1; c2tot = c2; c3tot = c3; c4tot = c4; end; else do; c1tot = c1tot+c1; c2tot = c2tot+c2; c3tot = c3tot+c3; c4tot = c4tot+c4; end; if last.wordmtch then output; drop c1 c2 c3 c4 words; rename c1tot = c1 c2tot = c2 c3tot = c3 c4tot = c4 wordmtch = words; run; Hope this helps,   Lee ... View more