topic Re: How do I remove some part of observation and keep the rest? in SAS Procedures

How do I remove some part of observation and keep the rest?

Sapkota — Wed, 16 Jul 2014 15:44:01 GMT

I have a dataset that has character part and the numeric part. I do not need the character part of the variable in the final analysis. How do I remove some part of the observation (i.e. the character part) and keep the rest (i.e. the numeric part).

For example in the dataset below, how do I program SAS so that 'cm' are removed from the 'Height' variable and only the numeric value is remaining. Same thing for Weight.

ID Age Height Weight

1001 15 163cm 140Lb

1002 21 180 145lb

1003 23 154cm 165

1004 19 132 134lb

1008 21 129cm 134lb

1009 14 113cm 123

1010 9 100 100

Any help in this matter will be highly appreciated.

Thanks

Sapkota

Re: How do I remove some part of observation and keep the rest?

Reeza — Wed, 16 Jul 2014 15:52:43 GMT

Look at the compress function. You can't change a character variable to numeric so you'll need to create a new variable. If you really need it to be the same variable name, rename the height variable before the conversion in the set statement.

height_num=compress(height, 'kd');

Re: How do I remove some part of observation and keep the rest?

slchen — Wed, 16 Jul 2014 15:55:42 GMT

height=compress(height,,'kd');

Re: How do I remove some part of observation and keep the rest?

stat_sas — Wed, 16 Jul 2014 18:19:55 GMT

To get height and weight as numeric variables, you need to use input function in addition to compress

data want;

set have;

height_num=input(compress(height,,'kd'),best.);

weight_num=input(compress(weight,,'kd'),best.);

run;

Re: How do I remove some part of observation and keep the rest?

Sapkota — Wed, 16 Jul 2014 19:08:56 GMT

Thank you for the helpful responses.

I have one related question. How do I delete the part of the numeric value from one observation which is entirely numeric. For instant in the following example, in the 'Height' variable, if I would like to delete the last two digits of all the observation (so that Height of ID 1001 would be12 instead of 1230 and Height of ID 1002 would be 14 instead of 1435), how do I do it?

ID Age Height

1001 10 1230

1002 12 1435

1003 13 1453

1004 15 1432

1005 11 1786

1006 12 1234

Any help will be highly appreciated. Thanks

Sapkota

Re: How do I remove some part of observation and keep the rest?

stat_sas — Wed, 16 Jul 2014 19:30:42 GMT

Try this for the desired output.

data want;

set have;

new_height=int(divide(height,100));

run;

Re: How do I remove some part of observation and keep the rest?

venkatnaveen — Thu, 17 Jul 2014 04:32:13 GMT

Why 'kd' used?

I am not getting..

Re: How do I remove some part of observation and keep the rest?

marty_ca — Thu, 17 Jul 2014 10:38:12 GMT

It means keep digit and seems to be a SAS option for the compress function. That is quite handy instead having to specify the actual characters.

Re: How do I remove some part of observation and keep the rest?

stat_sas — Thu, 17 Jul 2014 13:45:58 GMT

k and d are two modifiers in compress function that will keep digits in a string and throw away all the rest.