Hi, I am having SAS Snippet to directly export the any SAS dataset in .txt format based on the structure of that SAS dataset...You can find the attached file to solve your question...Hope it meets your requirements... This snippet i developed where i have not considered Delimeter facility so by default it will put space between the variables as delimeter... -Urvish ... View more

If you want row wise comparision (my output gives more insight) then you can simply eliminate the comparision which is not relevant for your business, from the final dataset... ... View more

Hi, If you want to find the out how many times a particular customer is moving to different retailer for different products and for specific time interval, then you need to include all possible mevements...Here in sample data, there are 3 retailers so you need to find out following: How Many times customer is moving from WALLMART to BESTBUY for different products How Many times customer is moving from WALLMART to FRYS for different products How Many times customer is moving from BESTBUY to FRYS for different products Here is my try, don't know it meets your requirement...But based on the business logic you mentioned in your very first post, i prepered the SAS Code as follow: proc sql;    create table want as    select a.consumer_id,a.retailer as from_retailer,           b.retailer as to_retailer,a.product_code,           a.shopping_date as from_date,b.shopping_date as to_date,           count(a.consumer_id) as FP    from have as a,have as b    where a.Consumer_ID = b.Consumer_ID    group by 1,2,3,4,5,6; quit; data want;   set want;   if from_retailer = to_retailer or   from_date GT to_date then delete; run; proc sort data = want;   by from_retailer to_retailer FP; run; data want;   set want;   by from_retailer to_retailer FP;   if last.from_retailer or last.to_retailer and last.FP then output; run; proc sort data = want(drop = product_code);    by from_date to_date FP; run; -Urvish ... View more

Hi, First, just add two variables in your TABEL1 as below: temp_field2 = field2; temp_field3 = field3; And use the following code to merge back with tabel1 to add CODE information... data both(drop = field2 field3 rename = (temp_field2 = field2 temp_field3 = field3));    merge one(in = a) two(in = b);    by field1 field2;    if a; run; -Urvish ... View more

Hi, I assumed first ID(that is ID of 1996 in this case) will be same to all different Firms... proc sort data = have;   by name; run; data temp(keep = id name);   set have;   by name;   if first.name then output; run; data want(drop = id);   retain year id unique_id name;   merge have temp(rename = (id = unique_id));   by name; run; -Urvish ... View more

Here, i think there is no need to use the PROC SQL if you can achive the same output by using Data Step... The below one is modefied code which achives the same output even if the dates are not in proper structure... If you think about the efficiency then Data Step is more efficient in this case as SAS Has to use summary functions in HAVING Clause in order to subset the group which of course leads the SAS to make extra pass through the data... proc sort data = one; by id date; run; proc sort data = two; by id date; run; data both;   merge one two(rename = (date = date1));   by id;   if first.id then output; run; -Urvish ... View more

How about the following one: data both;   merge one two(rename = (date = date1));   by id;   if first.id then output; run; -Urvish ... View more

Here is my try... data have;   input cat1 - cat3;   _temp1 = compress(cat1||cat2||cat3); /* Create Macro Variable if number of variables are more */   cards4; 1 0 0 0 1 0 0 1 0 1 1 1 1 0 1 ;;;; proc sort data = have;   by _temp1; run; data want(drop = _temp1);   set have;   by _temp1;   if first._temp1 EQ last._temp1 then unique_profiles + 1; run; If there are so many variables in your dataset then include all those variable names in to one macro variable... -Urvish ... View more

Based on my knowledge, it is not possible to apply multiple format to different rows for variable...You can read the variable containing different dates structure as charcter variable and can fix it as charcter variable... However, it will be good if you can share the purpose of your analysis by using this date variable in subsequent step... Usually, variable types creates conflict in appending and joining (merging) SAS Datasets... If your aim is structurize the same dates for all the rows then need to think out of box!!! -Urvish ... View more