BookmarkSubscribeRSS Feed
deleted_user
Not applicable
I need to concatenate two data sets using outer union corr. But these two datasets have different number of variables

WORK.READXTRAN has 196955 observations and 264 variables
WORK.READXTRAO has 1012170 observations and 245 variables

I'm trying to do:

proc sql;
Create table readxtra as
select *
from readxtraN
OUTER UNION CORRESPONDING
select *
from readxtraO;
quit;


I get an error: ERROR: Column 203 from the first contributor of OUTER UNION is not the same type as its counterpart

Please suggest how I should contenate these two datasets using proc sql?
4 REPLIES 4
Peter_C
Rhodochrosite | Level 12
you'll need to rename that column, and any others which have different datatypes for the same name.
The message indicates its number not its name, but PROC CONTENTS will provide the numbers.
christinagting0
Quartz | Level 8

Can someone elaborate on this? I am having the same issue, but I don't understand the solution here.

 

What do you mean rename the column? Just name the variable name? 

 

I have a number of these errors! Does that mean I have to rename every single column? What do I rename them to?

Patrick
Opal | Level 21

Run below code. That will hopefully explain it to you.

 

data have1;
  attrib 
    A length=8
    b length=8;
  A=1;
  B=1;
run;
data have2;
  attrib 
    A length= 8
    B length= $20;
  A=2;
  B='2';
run;

/* throws an error because column B is of different type in source tables  */
/* error indicates which column based on variable number causes the issue  */
/* running proc contents over source tables will return the variable names with attributs */
/*   so the problematic column is easily identified                        */
proc sql;
  create table demo1 as
  select *
  from have1
  outer union corresponding
  select *
  from have2
  ;
quit;

/* will work */
proc sql;
  create table demo1 as
  select a, b as b1
  from have1
  outer union corresponding
  select a, b as b2
  from have2
  ;
quit;
christinagting0
Quartz | Level 8

Thanks Patrick for taking the time to post the example. It really does help!

 

I have a question though. 

 

I notice that the correct proc sql code renames variables b to b1 from dataset have 1,  and then renames b to b2 from dataset 2. But for the merging of the two datasets I want b1 and b2 to be in the same column.

 

So for example:

 

data set have 1 has variables a and b

data set have 2 has variables a and b

 

final merged data set has variables a and b which contains the data points from variables a and b from dataset have 1 and have2.

 

is there a way to do this?

 

Thank you

SAS Innovate 2025: Register Now

Registration is now open for SAS Innovate 2025 , our biggest and most exciting global event of the year! Join us in Orlando, FL, May 6-9.
Sign up by Dec. 31 to get the 2024 rate of just $495.
Register now!

What is Bayesian Analysis?

Learn the difference between classical and Bayesian statistical approaches and see a few PROC examples to perform Bayesian analysis in this video.

Find more tutorials on the SAS Users YouTube channel.

SAS Training: Just a Click Away

 Ready to level-up your skills? Choose your own adventure.

Browse our catalog!

Discussion stats
  • 4 replies
  • 20451 views
  • 1 like
  • 4 in conversation