I got an awkward data selection problem.   data have; input x y1-y5; datalines; 816016.2     751112,abcklop     816015,wxyp     816016,klmno 140692.5     140672,aopk        140638,bfgrt       140682,uvgt       140652,kjnh      140692,kjnh 854222.3     854442,bnghy      854222,hhgi       821004,klop       821202,bmnh 216004.2     216004,klopi        250200,brtn        772551,jklop 853332.3     874089,kjlnm       816015,qwert      816014,mnbp       853332,ubny      740210,mnbp 161002.2     716022,gmbhj     161002,dmdn      810202,yupoi        816000,klhnp 450212.3     854001,mjkopl     441406,jkoprt     450212,kmplo       751101,weqrt       816018,kjmioot 858122.2    858112,brdt          857601,olpyt      857501,kmopyth     857401,mnkop   858122,rmnhg ; First six numbers in the row in column x matche with one of the first six numbers in the same row of column y1 through y5. But column y1 to y5 has both numbers and characters mixed, however the first six are always the numbers. I needed to select the one value (both the six numbers and attached characters) from y1 to y5 that matches the six numbers in column x and drop the rest of the unwanted values. For example, Data want; 816016.2      816016,klmno 140692.5      140692,kjnh 854222.3      854222,hhgi 216004.2      216004,klopi 853332.3      853332,ubny 161002.2      161002,dmdn 450212.3      450212,kmplo 858122.2      858122,rmnhg I would appreciate very much if some expert in the community can help me to get the result.   Thank you all, SkSap ... View more

Got a question. Data have; Input X1-X10; cards; 7.2    8.1    8.2    9.1 1.1   1.4     5.2     2.3    7.1    8.1    1.5 8.1    8.1    8.3    8.2    8.1 2.1    2.2    2.5    7.3. 7.1    7.2    8.1    8.0    8.3    8.1    9.1 ..........; (thousands of rows) I want to reorder the rows in descending order according to the number after the decimal. Front number doesn't matter as long as numbers after the decimal are  are in descending order. But, the numbers before and after the decimal should be together. In other words, Data want; 7.2    8.2    8.1    9.1 1.5    1.4    2.3    5.2     7.1    8.1    1.1 8.3    8.2    8.1    8.1    8.1 2.5    7.3    2.2    2.1 8.3    7.2    7.1     8.1    9.1   8.1   8.0 ..................,   I thank you very much if any one of the expert out there could help me to get the data I wanted.   ... View more

Got a data set merging problem: Data Have1; a1 a2 a3 . 4 3 1 . . 2 3 4 4 9 3 9 5 2   data have2; a1 b1 b2 1 6 5 2 4 . 9 2 3 12 1 1 I want to merge data Have1 and Have2 in such a way that if the values of the variable a1 in Have1 match with values of a1 in Have2 then keep otherwise drop. For example I want the new set like: a1 a2  a3 b1 b2 1 . . 6 5 2 3 4 4 . 9 5 2 2 3   I would appreciate very much if any expert out there help me out to get the result.   Thank you very much. Sijansap           ... View more

Dear experts, I have 3 columns of 3-digit data with about 30k rows. I want to concatenate and put commas between them to make a single character variable and use PROC FREQ to get the count of each of the similar combinations. The problem is I want to treat, say ABC is same as ACB, or BAC, etc. (all 6 possible combination same) as same entry and if all 6 combinations are present in the data then PROC FREQ should yield ABC with freq count 6. Also, if there are only two A and B the treat AB and BA same. The data looks like:   DATA HAVE; input A B C; D= CATT(of A, B, C); cards; 801 905 823 903 801 803 905 801 823 823 801  . 823 905 801 801 823 . ...; Now D will have 1st, 3rd, and 5th row will have (801,905,823), (905,801,823) and (823,905,801), I would like PROC FREQ to yield (801, 823, 905) with frequency 3, And, similarly from 4th and 6th row (801,823) with frequency 2.   I will appreciate any suggestions. Thank you all. Sijansap ... View more

  Dear Experts, I got a situation. It is about body part injury and admission data. H being head, C being chest, A abdomen etc. 6 different body parts, and 1, 2, 3, and 4 being severity codes 1 being mild and 4 being severe. Each row is a case (or say admission to health center). These letter and number codes are combined together and has 20,000+ cases. I would like to select one letter code with highest severity code (number) and make one column say MaxBI (body part with maximum injury). Some of the cases don't have all 6 body injuries. Following is what I have: Data have; input $ BI1-BI6; cards; H1   C2   F4   A2   G1   E2 H3   C1   F2   A4   G2   E1        C2   F3          G4   E2 H4          F1                  E3 ... ; Data Want: /*with MaxBI*/ H1   C2   F4   A2   G1   E2         F4 H2   C1   F2   A3   G2   E1         A3        C2   F1          G3   E2         G3 H2          F1                  E1         H2    ... I could split the letters and numbers from the first six columns using Substring function and take the Max of the substringed values, but I would not know which body part it came from. That is where I got stuck. I would appreciate very much for your kind help. Thank you. Sijansap ... View more