Delete empty rows in SAS - sas

Question

I have data set that contains empty cells. It looks like
Year Volume ID
2000 999 LSE
2001 . LSE
. 555 LSE
2008 . NYSE
2010 1099 NYSE
I need to delete the row that contains empty cells. The output should look like this
Year Volume ID
2000 999 LSE
2000 1099 NYSE
I tried following code
data test;
set data;
if volume = " . " then delete;
if year= " . " then delete;
run;
But output file has 0 observations and SAS gives me
NOTE: Character values have been converted to numeric values at the
places given by (Line):(Column).
Also I tried
options missing = ' ';
data test;
set data;
if missing(cats(of _all_)) then delete;
run;
But its not working as well.
I just want to delete the rows with empty cells.
Anyone can help me to solve this issue ? Thanks in advance !!!

Answer 1

Options Missing only affects how things are printed or converted when going numeric -> character. In this case you have numerics, so it accomplishes nothing.
Your first code sample is mostly correct- at least, when I try it, it works. " . " is not really right, but it will convert (as the note says) to missing since none of those characters are a number.
The proper way to do this is one of the two:
data have;
input Year Volume ID $;
datalines;
2000 999 LSE
2001 . LSE
. 555 LSE
2008 . NYSE
2010 1099 NYSE
;;;;
run;
data want;
set have;
if year = . then delete;
if volume = . then delete;
run;
or
data want;
set have;
if missing(year) then delete;
if missing(volume) then delete;
run;
missing returns true if the variable is missing (which includes 28 total values, but . is the most common).
A better way to do more than one is to use the nmiss or cmiss functions (nmiss for numbers, cmiss for character or mixed type).
data want;
set have;
if nmiss(year,volume) = 0;
run;
That will return the number of missing values, which you can then test for whatever value you are looking for (in this case, zero values). You could even do:
data want;
set have;
if nmiss(of _NUMERIC_) = 0;
run;
where _NUMERIC_ is all numeric variables. (of is needed for variable lists like this to tell SAS to expect a list.)
Your second doesn't work, by the way, because it's catting the ID variable together with the others. You could have seen this by looking at the value of that cats (ie, assign it to a variable). You could have said
if cats(of _all_) = ID then delete;
but as several of us have shown that's probably inferior to the simpler solutions using nmiss.

Answer 2

You can just use a subsetting if nmiss() by checking the variables that must be populated:
data test;
set data;
if nmiss(year,volume)=0 ;
run;
Edit: This works if year and volume is numeric, if it is string, you can use the cmiss() function.

Answer 3

Don't use quotes with numeric variables, e.g.:
if volume = . then delete;
Other option that works for either character or numeric:
if missing(volume) then delete;

Answer 4

You could use a where clause in the set statement here as well:
data new_dataset;
set old_dataset (where = (volume is not missing or year is not missing));
run;
I always enjoy using the is not missing syntax because it seems too much like writing normal English to work