topic Re: How to recode a categorical variable in SAS Programming

How to recode a categorical variable

SSin — Sun, 02 Jun 2019 08:46:23 GMT

I have a data set which contains Character variable "Housing_type". The response to that variables are "3-room", "4-room", "5-room", "Condominium", "Landed property" and "Other".

I would like to recode 3-room, 4-room and 5-room into "HDB" and Condominium, Landed Property and Other into "Non-HDB".

How may I do so please?

Re: How to recode a categorical variable

Astounding — Sun, 02 Jun 2019 13:11:11 GMT

Have you tried using IF THEN statements in a DATA step? You will be using them virtually every day, so it's a good time to learn them. Give it a shot and post the results if you have any difficulty.

Re: How to recode a categorical variable

Reeza — Sun, 02 Jun 2019 23:08:05 GMT

You can use IF/THEN to recode variables. The general structure is as follows:

length new_housing_type $12.;

if housing_type  in ('3-room'  '4-room' '5-room') then new_housing_type = 'HBD';
else if housing_type in ("Condo") then new_housing_type = 'Condo';
else new_housing_type = 'Other';

You'll need to modify it to match your exact cases. Note these are case sensitive comparisons.

@SSin wrote:

I have a data set which contains Character variable "Housing_type". The response to that variables are "3-room", "4-room", "5-room", "Condominium", "Landed property" and "Other".

I would like to recode 3-room, 4-room and 5-room into "HDB" and Condominium, Landed Property and Other into "Non-HDB".

How may I do so please?

Re: How to recode a categorical variable

ballardw — Mon, 03 Jun 2019 15:04:36 GMT

@SSin wrote:

I have a data set which contains Character variable "Housing_type". The response to that variables are "3-room", "4-room", "5-room", "Condominium", "Landed property" and "Other".

I would like to recode 3-room, 4-room and 5-room into "HDB" and Condominium, Landed Property and Other into "Non-HDB".

How may I do so please?

Another approach is to use a custom format, which if you do this frequently may be a better solution. Example:

Proc format library=work;
value $ht
"3-room", "4-room", "5-room" = 'HDB'
"Condominium", "Landed property" , "Other" = 'Non-HDB'
;

data want;
  set have;
  ht = put(housing_type,$ht.);
run;

Or try

Proc Print data=have;
   format housing_type $ht.;
run;

An advantage of a custom format is that if you find that you need to add an item, such as possibly "6-room" then you only need to add it to the format definition and rerun the proc format code. No need to add extra if/then/else.

If using the format directly with the variable instead of creating an additional variable then you have the option to use the groups dynamically in most analysis, reporting or graphing procedures.

Such as:

Proc format library=work;
value $ht
"3-room", "4-room", "5-room" = 'HDB'
"Condominium", "Landed property" , "Other" = 'Non-HDB'
;
value $ht_3lvl
"2-room","3-room"='Small HDB' 
"4-room", "5-room" = 'Large HDB'
"Condominium", "Landed property" , "Other" = 'Non-HDB'
;
run;
Proc freq data=have;
   tables housing_type;
   format housing_type $ht_3lvl.;
run;

I of course have no idea if "Small" or "Large" actually apply to your need, that is just an example.

Since your data likely does not have any 2-room from your description the "Small" would just be 3-room.

Re: How to recode a categorical variable

SSin — Tue, 04 Jun 2019 03:47:59 GMT

Many thanks!!

Re: How to recode a categorical variable

SSin — Tue, 04 Jun 2019 03:48:29 GMT

Many thanks!!

Re: How to recode a categorical variable

SSin — Tue, 04 Jun 2019 03:52:43 GMT

I used If/then. It is like that If/then statement totally does not work. Normally, if and then suppose to go blue if the statement works. But they remains faded and when I run it nothing happen. That is why. I am using SAS university edition for the first time.

Re: How to recode a categorical variable

SSin — Tue, 04 Jun 2019 03:54:12 GMT

I used If/then. It is like that If/then statement totally does not work. Normally, if and then suppose to go blue if the statement works. But they remains faded and when I run it nothing happen. Do you have any idea why it is not working at all? I am using SAS university edition for the first time.

Re: How to recode a categorical variable

Astounding — Tue, 04 Jun 2019 14:09:02 GMT

I'm squinting, but I still can't see your program. If you post the entire DATA step that you tried, I would be happy to respond.