topic Re: change from character to numeric for a column/variable in SAS Data Science

change from character to numeric for a column/variable

mei — Fri, 22 Jun 2012 02:24:03 GMT

Dear Sir,

I wish to merge several files with similar variables but certain variables are defined as character in one file but as numeric in another file.

I need to standardize all these variables from character to numeric, ie Yes=1, No/others=0.

How do i do it?

Pls refer to file attached:

Eg: DualClass variable -in character- yes or no need to convert to yes=1, no=0.

other variables are BLANKCHECK CBOARD LSPMT

Pls kindly assist. Thanks a lot.

Re: change from character to numeric for a column/variable

mei — Fri, 22 Jun 2012 02:33:08 GMT

can i use input function, or should i create a dummy variable first to 1 and 0 say name it as DualClassDummy. Then i will drop the original variable DualClass and rename dualclass dummy as dual class? (so that i can do further merging with the rest of the files?)

Thanks

Re: change from character to numeric for a column/variable

Linlin — Fri, 22 Jun 2012 13:02:03 GMT

try using the code below as an example to write your own code:

data have;

input DualClass $;

cards;

yes

YES

;

data have(drop=DualClass rename=(_DualClass=DualClass));

set have;

if upcase(compress(DualClass))='YES' then _DualClass=1;

if upcase(compress(DualClass))='NO' then _DualClass=0;

run;

proc print;run;

Re: change from character to numeric for a column/variable

MikeZdeb — Fri, 22 Jun 2012 14:22:39 GMT

hi ... what do you want is a value is MISSING, another MISSING or consider it a NO (using two of your yes/no variables)

this makes MISSING = MISSING ...

data have;

input (dualclass carveout) (:$3.) ;

datalines;

yes .

no NO

YES .

. YES

NO yes

;

proc format;

invalue yn(upcase) 'YES' = 1 'NO' = 0 other = . ;

run;

data want;

set have;

_dualclass = input(dualclass,yn.);

_carveout = input(carveout,yn.);

run;

with a lot of yes/no variables, this might save some keystrokes ...

data want;

set have;

array old(*) <list all the y/n variables>;

array new(*) <list all the new variables ... same order as old>;

do _n_=1 to dim(old);

new(_n_) = input(old(_n_),yn.);

end;

run;