topic Re: Variable has been defined as both character and numeric in SAS Programming

Variable has been defined as both character and numeric

Venki — Tue, 22 Jan 2013 01:42:42 GMT

I am trying to combine 8 datasets of different patient data captured in 6 month intervals in 2 different hospitals. So, I have to combine those 8 datasets so that I have one nice dataset that I could work on.

When I try to combine using

data Combined; set fullextract1 fullextract2 fullextract3 fullextract4 fullextract5 fullextract6 fullextract7 fullextract8;

run;

I get multiple variables have been defined both as character and numeric like this:

ERROR: Variable DOB has been defined as both character and numeric.

So, I tried to put this in data statement for all involved variables in the individual datasets (basically trying to convert to character variables):

DOBch= put (DOB, $7.);

drop DOB; rename DOBch=DOB;

But, still when finally try to combine using above combine statement, I still get these error statements. Please help.

Re: Variable has been defined as both character and numeric

PGStats — Tue, 22 Jan 2013 02:14:59 GMT

Try to do your conversions this way :

data fullExtract1;

set fullExtract1(rename=(DOB=DOBnum));

DOB = put(DOBnum, 7.);

drop DOBnum;

run;

Re: Variable has been defined as both character and numeric

Venki — Thu, 24 Jan 2013 04:34:16 GMT

Thanks a lot !!!!

It worked very well...Much appreciated..Venki

Re: Variable has been defined as both character and numeric

_maldini_ — Tue, 29 Mar 2016 21:47:35 GMT

@PGStats

Just so I'm clear...The guidance is to convert the variables in each individual dataset before joining, correct?

Re: Variable has been defined as both character and numeric

PGStats — Tue, 29 Mar 2016 22:08:39 GMT

You can convert as you join. RENAME the wrongly typed variables in the SET statement, convert them in the code, assign the result to the properly named variables, and drop the renamed variable. You can do for multiple datasets the same as I proposed in this old post for a single dataset

Re: Variable has been defined as both character and numeric

NKormanik — Thu, 19 May 2016 05:32:47 GMT

How would I combine two cases of your solution?

data nicholas._21305_21501 ;
set nicholas._21305_21501 (rename=(i1_Value=i1_Value_Character)) ;
i1_Value = put(i1_Value_Character, 1.) ;
drop i1_Value_Character ;
run ;

data nicholas._21305_21501 ;
set nicholas._21305_21501 (rename=(i2_Value=i2_Value_Character)) ;
i2_Value = put(i2_Value_Character, 1.) ;
drop i2_Value_Character ;
run ;

Re: Variable has been defined as both character and numeric

PGStats — Thu, 19 May 2016 14:24:11 GMT

data nicholas._21305_21501 ;
set nicholas._21305_21501 (rename=(i1_Value=i1_Value_Character i2_Value=i2_Value_Character)) ;
i1_Value = put(i1_Value_Character, 1.) ;

i2_Value = put(i2_Value_Character, 1.) ;
drop i1_Value_Character i2_Value_Character ;
run ;

and so on, for all wrong typed variables.

Re: Variable has been defined as both character and numeric

hovliza — Mon, 23 Oct 2017 09:40:33 GMT

I had the same problem, but when I do this, in my final table there are no values for "p_no". There is a column, but the values are missing.. What can I do about it? Something with the ''7.''? What does that mean?

DATA fullExtract1 (rename=(respno=p_nonum));
SET fullExtract1;
p_no = put(p_nonum, 7.);
DROP p_nonum;
RUN;

Re: Variable has been defined as both character and numeric

PGStats — Mon, 23 Oct 2017 17:43:38 GMT

Start a new forum topic, post the log and an example of the value of respno.