topic Re: Null columns in SAS Programming

Null columns

HeatherNewton — Sat, 10 Dec 2022 02:07:40 GMT

If i have series of program that produce series of tables containing production data, instead of masking some sensitive fields, if i just make these data null, e.g. have all names become null in the same column, is all programs still run as before given this name field is not key to any join table or a factors in any important calculation

Re: Null columns

Tom — Sat, 10 Dec 2022 03:44:11 GMT

@HeatherNewton wrote:

If i have series of program that produce series of tables containing production data, instead of masking some sensitive fields, if i just make these data null, e.g. have all names become null in the same column, is all programs still run as before given this name field is not key to any join table or a factors in any important calculation

Is there a question there somewhere?

If you want to make a copy of the data and set some variables to missing then just use the CALL MISSING() function.

For example if you wanted to blank out NAME and AGE from SASHELP.CLASS you could do this:

data want;
  set sashelp.class;
  call missing(of name age);
run;

Re: Null columns

SASKiwi — Sat, 10 Dec 2022 03:56:40 GMT

BTW NULL is a database concept where database columns store special null values which are not the same as missing or blank values. SAS columns only contain missing values, not null.

Re: Null columns

LinusH — Mon, 12 Dec 2022 12:22:45 GMT

From syntax perspective, yes.

But what is the point of keeping them at all if they are not contributing to the result?

Re: Null columns

ballardw — Mon, 12 Dec 2022 16:50:13 GMT

@HeatherNewton wrote:

If i have series of program that produce series of tables containing production data, instead of masking some sensitive fields, if i just make these data null, e.g. have all names become null in the same column, is all programs still run as before given this name field is not key to any join table or a factors in any important calculation

If I understand your question the answer could be "maybe". Depends on how you intrepret or specify "important".

If you are using the variables as part of By group or Class statement then all missing values within a By group/ Class group with multiple variables will be the same "value" with the non-missing values. Depending on the code actually used missing values may be included or know (look for option MISSING in some procedures). If procedures us Class statements then typically missing values of the Class statements will result in records being excluded from use in the procedure.

As is often the case, you might be better off just trying your process with and without the values in the data set and compare the results.