Delete observations which include certain text in one of many variables

[jimbarbour]

It would be helpful if you post some data -- even if it's just mocked up data.  

One question I have is:  Can "Toyota" or "Honda" appear within other text or is the brand name of the car in a column by itself?

For example, would your data look more like the "Brand" column below or would it look more like the "Car_Info" column below?

Brand	Car_Info
Toyota	2014 Toyota Camry

If "Toyota" can appear anywhere in the column, then we need to take a different approach. 

Jim

[mgrasmussen]

Dear Jim

Thanks for the input.

I realize that I should have specified this detail. The data is as Brand, but at some point I will need to find similar solutions for variables which are organised as 'Car_Info'. I believe another person posted code which would work in such a scenario (with 'Car_info').

[ErikLund_Jensen]

Hi @mgrasmussen 

Here is a slightly modified version of @tarheel13 's code. It uses an automatic array _characters_ containing all defined character variables (a smart SAS feature). The index function + lowcase identifies all occurences of Toyota or Honda in all character strings regardless of case or other content in the text.

data mydata;
	set sashelp.cars;
	array cars $ _character_;
	flag='N';
	do i=1 to dim(cars);
		if index(lowcase(cars[i]),'toyota') or index(lowcase(cars[i]),'honda') then flag='Y';
	end;
	if flag='Y' then delete;
run;

[mgrasmussen]

Dear Erik

Thanks. These are great features and would work in my case.

What is the purpose of the lowcase function? I believe I understand what it does, but why is it necessary here? I would suspect that this specification is related to the fact that you write 'toyota' and 'honda' within the code in brackets. 

[tarheel13]

Lowcase function is more defensive programming in case your variables were not all in the same case. Using this function will capture Toyota or Honda regardless of case