When you ask "why" we need either the data files used, which may have problems with sharing, and the log of the step you question. So go to the log where you get the empty table, copy the CODE and all notes or messages from that step. Then on the forum open a text box using the </> icon above the message window and paste the text.   Most likely cause from the code I see is that since you use BY _ALL_ in the Merge then one or more variables does not have the same value in both sets. In fact since you sorted and sub-setted the data on Review_count variable values I would be extremely surprised if the "most reviewed" and "least reviewed" have the same number of reviews. So they would not merge onto the same observation. Pick actual identifying variables for a BY. Do not include variables like "counts" unless you expect most of them to match. And if you have two different variables for the count, implied but not shown, then don't expect them to match very often.   Something else,  you may want to rename the Review_count variable from one or both of the sets so that you have two variables to actually compare. I do not see where Review_count_1 or Review_count_2 are assigned/created from I have to assume from Review_count in Qed.Products. ... View more

You can streamline your process a lot: data availableproducts unavailableproducts ; set qed.products; rating_num = input(rating,best32.); grams_num = input(grams,best32.); ptice_num = input(price,best32.); drop rating grams price; rename rating_num=rating grams_num=grams price_num=price ; select (available); when ("T") output availableproducts; when ("F") output unavailableproducts; end; run; proc means data=availableproducts; var grams price rating; output out=availablestats mean=avg_grams avg_price avg_rating ; run; But you should first ask yourself why the conversions are needed in the first place. Numbers stored in character variables point to a faulty import process which brings the data into SAS, e.g. using a bad medium (Excel files) or PROC IMPORT. ... View more

A single dataset would not be stored in 6 separate worksheets.  That would be 6 separate datasets.  If the data is related you might be able to combine them.   How did you convert the workbook into SAS datasets?  Did the worksheet with the 1,069 rows yield a dataset with 1,068 observations (assuming the first row in the worksheet has column headers that SAS can use to guess what names you want to use to refer to the variables it creates).   The usual reason why only 30 out of 1,068 observations participate in an analysis is that observations with missing values for one or more of the variables are excluded.  How many variables are in the dataset you used for the analysis?  How many of the variables did you use in the analysis?  How many missing values do the variables have? ... View more