Hi: Here's an example using SASHELP.CLASS.
The keywords for count and percent with PROC REPORT are N and PCTN. PCTN gives the percent of the grand total. SASHELP.CLASS has 19 observations so you can see that each percent is the number in the cell divided by 19. Then the percent9.2 format causes the multiply by 100 to make the percent value.
If you have 2 "sub groups" named SUB-01 and SUB-02, then you need to have an ACROSS item in your PROC REPORT COLUMN statement. You can nest N and PCTN under the ACROSS item, in this case the SUB variable, as shown below:
Since SASHELP.CLASS did not have a variable called SUB, I made one. SUB-01 group still has 19 observations, but SUB-02 value only has 13 observations. The percents for SUB-01 will be calculated by dividing by 19 and the percents for SUB-02 will be calculated by dividing by 13.
As you can see, the default behavior of PROC REPORT is to give 1 statistic per column. You CAN manipulate the values to make a new character variable that combines N and PCTN into 1 value like this 19 (100.00%), but that is something you would do AFTER you get a program similar to the above program working and that is the technique I show using COMPUTE blocks in my Creating Complex Reports paper that I referenced in my previous posting.
Cynthia
Hi:
There is always the possibility to get output the way you want -- however, based on what you've posted, I don't think you can create the desired output from the input data you show.
The challenge between what you show and what you want is that there's not really enough information or enough context for what you're trying to do to provide any suggestions. Here are my thoughts:
Your input data seems to be a jumble of State (TAMIL NADU, KARNATAKA and ANDHRA) combined with cities in one column. But I don't see, given that there's only 1 column in your data, how you can determine which cities go under which states in the output file.
As your input data stands, PROC SORT, PROC SQL, PROC REPORT, and PROC TABULATE will NOT give you what you want with only 1 column and the values jumbled together as they are. Perhaps if your data had a different structure or some kind of other identifier it would be possible. The only way I can see to produce your desired output is for you to manually associate the cities with each state.
Also, this seems to be a new question, and unrelated to the previous postings in this thread. In the future, it's probably better to open a new posting for a new question.
Cynthia
Learn the difference between classical and Bayesian statistical approaches and see a few PROC examples to perform Bayesian analysis in this video.
Find more tutorials on the SAS Users YouTube channel.
-
19706 views
-
3 likes
-
-
