topic Re: How do I use Arrays and DO loop to create new variables? in New SAS User

How do I use Arrays and DO loop to create new variables?

abrice520 — Mon, 26 Oct 2020 03:01:07 GMT

I'm struggling with a homework question for my SAS class. Using the provided data set "Mice" which contains the weight in ounces of 8 mice, I have to use arrays and a DO loop to create 8 new variables named "Diff1" to "Diff8" that contain the difference in weight between each mouse and the average weight of all 8 mice. I also have to calculate the average weight in the array.

Here is the provided Data:

DATA WORK.Mice;
	INFILE	DATALINES;
	INPUT	WtOz1 - WtOz8	@@;
	DATALINES;
0.57 0.69 0.87 0.86 0.83 0.63 0.68 0.59
;

Here is my attempt at the question following an example problem. I think I got the right values but they appear as "WtOz1" and so on rather than "Diff1" and I know there's a better way than to just rename the variables. I also don't know how to calculate the average within the array.

DATA MiceMean;
	SET Work.Mice;
	MeanW= (MEAN(OF WtOz1-WtOz8));
	ARRAY Diff{*} WtOz1 - WtOz8;
	DO i= 1 TO DIM(Diff);
	Diff{i}= MeanW-Diff{i};
	Drop i ;
	END;
	RUN;

Thank you!

Amanda

Re: How do I use Arrays and DO loop to create new variables?

jimbarbour — Mon, 26 Oct 2020 03:22:58 GMT

@abrice520

Amanda,

Everything you're doing is basically correct. You're saying MEAN which is the correct function, and you're using "of" which is the correct way to refer to multiple variables or an array. You're using a variable range, WtOz1 - WtOz2, instead of an array reference. An array reference for an array named WtOzs would be WtOzs[*].

Here's one possible way to code this:

DATA WORK.Mice;
	DROP	_:;
	INFILE	DATALINES;
	INPUT	WtOz1 - WtOz8	@@;

	ARRAY	WtOzs	[*]	WtOz1	-	WtOz8;
	ARRAY	DiffOzs	[*]	DiffOz1	-	DiffOz8;

	Avg_WtOz		=	MEAN(of	WtOzs[*]);

	DO	_i			=	1	TO	DIM(WtOzs);
		DiffOzs[_i]	=	Avg_Wtoz	-	WtOzs[_i];
	END;

DATALINES;
0.57 0.69 0.87 0.86 0.83 0.63 0.68 0.59
;
RUN;

Notice however that I have two arrays. One array is for the values read in, the weights. The other is for the calculation results, the differences. I wouldn't try to do this in one array although you could. If you don't want the weights in your results, then you could just code: DROP WtOz1 - WtOz2.

Since my two arrays are essentially mirror images of one another, a single subscript, _i, can be used to access both. The arrays function in parallel.

Jim

Re: How do I use Arrays and DO loop to create new variables?

ballardw — Mon, 26 Oct 2020 03:19:46 GMT

It isn't an "and" type question. You create new variables, one way at least, by declaring them in an array definition.

A statement like

Array diff(8);

will create 8 variables named diff1, diff2, and so forth.

You would assign values to each indexed variable inside the do loop though.

I would suggest NOT replacing the values in your weight variables so you can check results.

It might make sense to create a second array of those though, so you are pretty close:

DATA MiceMean;
	SET Work.Mice;
	MeanW= (MEAN(OF WtOz1-WtOz8));
	ARRAY wt{*} WtOz1 - WtOz8;
        array diff(8);
	DO i= 1 TO DIM(Diff);
	   Diff{i}= MeanW-wt{i};
	Drop i ;
	END;
RUN;

It is very common to use parallel arrays. Just make sure that they all reference things in the expected order which can be tricky when using less than optimal names for variables.

Re: How do I use Arrays and DO loop to create new variables?

abrice520 — Mon, 26 Oct 2020 16:32:21 GMT

Thank you so much!!