Originally used codes to create a macro variable: %let class = A B C; %global class; Now tried to use Prompt Manager to achieve the same result above: Users will be required to select from the list. If he/she selects B & C: Then the value for the macro variable 'class', I hope, is "B C". Clearly the Log shows: 8 %LET class0 = %nrstr(2); 9 %LET class1 = %nrstr(B); 10 %LET class2 = %nrstr(C); 11 %LET class_count = %nrstr(2); 12 %LET class = %nrstr(B); which is not what I need.   How to fix it? ... View more

Version: SAS Enterprise Guide 7.1   Programmers have defined a bunch of global macro variables. e.g. %let example = 4; The question is how to check whether/where the defined macro variables are used in other Programs (even in different process flows). ... View more

Original dataset: Type Class ... Total ... Category 1. Need to move the column 'Total' to the end without changing the order for the rest of columns. The result looks like: Type Class ... Category Total 2. Need to move the column 'Total' to the beginning without changing the order for the rest of columns. Total Type Class ... Category Requirements: Hard-coding like retain ColA ColB ... ColZ Total; wouldn't work because we wouldn't know the names for the rest of the columns or how many columns there are (they are dynamic).   Thank you. ... View more

Users define a couple of macro variables (input): %let category = A B; %let class = apple banana; %let column = index character; The generated table/dataset (output) would look like: category class index character A apple . . A banana . . B apple . . B banana . . As you can see, for column 1 and column 2, they basically go through values of the macro variables, with macro variables' names as the column names. For column 3 and column 4, blanks, with macro variable's value as the column names (column 'index' has numeric values; column 'character' has char.)   How to achieve it? ... View more

The orginal dataset has columns like: Time A_1 A_2 B_1 B_2 C_1 and all the columns except 'Time' have numeric values.   Need the codes that can automatically add new columns: in this case,  A_All B_All C_All and A_All = A_1 + A_2 B_All = B_1 + B_2 C_All = C_1   The problem is that, the column names of the original dataset, instead of 'A_:' 'B_:' 'C_:', can be anything (e.g. 'Apple_:', 'Test_:' etc.;  the new columns would accordingly have names like 'Apple_All', 'Test_All'). In addition, there could be any number of 'A_:' columns: e.g. 'A_1-A_7' Hard coding like: A_All = sum(of A_1-A_2); wouldn't work therefore.   I'm expecting codes that achieve the following objectives: 1. Read column names from the original dataset - have; 2.1 for columns whose names have a format like 'alphabetic characters + underscore + numbers': e.g. Type_3 2.2 identify how many columns there are whose names have the same alphabetic characters: e.g. 'Type' 3. create new columns whose names are 'same alphabetic characters + underscore + "All"': e.g. Type_All 4. the value for the new column is therefore: e.g. sum(of Type_1-Type_&NumColType).   To make the question easier, the numeric part of the column names always starts from '1', and increases by 1. e.g. for 'Type_:' columns, it would be Type_1 Type_2 Type_3. ... View more

/* Users specify a macro variable */ %let Name = 0 test All %let Num_Name = %sysfunc(countw(&Name)); /* which is 3 */ %do _z=1 %to &Num_Name; %let File_Name = %scan(&Name,&_z); /* datasets have names like */ lib.data_&File_Name /* in this case, the dataset names would be lib.data_0 lib.data_test lib.data_All */ These three datasets have the same columns:  /* e.g. one dataset would look like */ Type _1 _2 _3 Order A 0 0 0 1 I would like to concentrate these datasets, and create a final dataset: lib.data_final (without hardcoding it) and then rows arranged in ascending order based on the column 'Order'.   Tried using a loop: %do _z=1 %to &Num_Name; %let File_Name = %scan(&Name,&_z); data lib.want; set lib.want lib.data_&File_Name; However reported an error: lib.want.data does not exist. ... View more

/* dataset */ _NAME_ new1 new2 new3 new4 Time_0 1 2 3 4 Time_1 5 6 7 8 Time_2 9 10 11 12 Time_3 13 14 15 16 I have a dataset, and need to use a two-dimensional array - temprg{i,j} for some data manipulation. how to assign value to this array?   Most online SAS tutorials use Datalines, but since I have much more rows and columns, I'm looking for another way to assign the initial values. e.g. temprg{2,3} would be 7. temprg{4,2} would be 14.   Thanks. ... View more

/* data have */ Period Time_0 Time_1 Time_2 Time_3 Period_1 0 4 7 9 Period_2 1 5 8 . Period_3 2 6 . . Period_4 3 . . . /* transpose first */ Time Period_1 Period_2 Period_3 Period_4 Time_0 0 1 2 3 Time_1 4 5 6 . Time_2 7 8 . . Time_3 9 . . . /* then mirror */ Time Period_4 Period_3 Period_2 Period_1 Time_0 3 2 1 0 Time_1 . 6 5 4 Time_2 . . 8 7 Time_3 . . . 9 /* lastly calculate */ if CountNum = 1: Time Result Time_1 3 /* 6/2 */ Time_2 1.6 /* 8/5 */ Time_3 1.29 /* 9/7 */ if CountNum = 2: Time Result Time_1 3.67 /* (5+6)/(1+2) */ Time_2 1.67 /* (7+8)/(4+5) */ Time_3 1.29 /* 9/7 */ if CountNum = 3: Time Result Time_1 5 /* (4+5+6)/(0+1+2) */ Time_2 1.67 /* (7+8)/(4+5) */ Time_3 1.29 /* 9/7 */   Basically sum up the same number (&CountNum) of numbers, and then it is divided by the sum of numbers from the same columns but one row above.   Thanks.   ... View more