topic Re: data step question: to tell if a term is belonged to a term_group in SAS Programming

data step question: to tell if a term is belonged to a term_group

Lulus — Thu, 27 Jun 2013 15:43:23 GMT

I am trying to do this to match an item (apple, green_onion, yam, orange...... ) with its own catgory.

Priror to it, I have tables that lists items in each group.

The two resouce tables are:

Vegie_table with two columns: item(50+) group (all values='vegie')

Fruit_table with two columns: items(50+) group(all values='frt')

drink_table with two columns: items(1000+) group(all values='drk')

My table look like this

Header 1	food	total price
name1	orange	1.5
name2	onion	2
name2	bean	2
name3	potato	2.8

I would like to create another column ----- food_type which indicates the type of the name in the food column.

I tried to use

if food is in (&itemlist.) then food_type=' xx ;

where itemlist is combined by vegie_table and fruit_table

but the itemlist I obtained has no quotes seperate them so I can not do it this way. Imagine I need to do this with my drink_list as well.

this is probaaly like SAS version of vlookup....

Please let me know how to do it with what I want. Thank you all~

Re: data step question: to tell if a term is belonged to a term_group

Reeza — Thu, 27 Jun 2013 15:55:53 GMT

If its a version of vlookup, look at proc format and put statements or merges.

You need to provide more detail otherwise.

See here:

Re: data step question: to tell if a term is belonged to a term_group

ballardw — Thu, 27 Jun 2013 16:32:50 GMT

Possibly

if index("&itemlist",food) >0 then food_type=' xx';

might work

Re: data step question: to tell if a term is belonged to a term_group

Tom — Thu, 27 Jun 2013 20:51:30 GMT

You should be able to convert your lookup tables into a format.

I would combine the three tables into one and use it as the CNTLIN

Vegie_table with two columns: item(50+)  group (all values='vegie')
Fruit_table with two columns: items(50+)  group(all values='frt')
drink_table with two columns: items(1000+)  group(all values='drk')

data formats ;

set vegie_table fruit_table drink_table ;

by item;

retain fmtname 'FOOD' type 'C' ;

rename item=start group=label;

run;

proc format cntlin=formats;

run;

....

food_type = put(food,$food.);