Date:         Thu, 4 May 2000 13:41:41 -0400
Reply-To:     Richard DeVenezia <radevenz@IX.NETCOM.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Richard DeVenezia <radevenz@IX.NETCOM.COM>
Organization: MindSpring Enterprises
Subject:      Re: Proc Compare

Sharavi:

Are you trying to do a conditional merge, in such that, you will replace the
values from XXXX with the values from YYYY only if a certain conditional
logic is met with regards to 'concordance' ?

e.g.
* XXXX and YYYY have the same variables;
data new;
  merge XXXX YYYY(rename=(long list of renames for 80 variables);
  if (complicated conditional logic comparing variable values from XXXX with
renamed variable values from YYYY) then do;
    <assign XXXX variables values of corresponding renamed YYYY variables>
 end;
run;

I suggest you create a table or algorithm to create a table that contains
the original variables names and the 'new' variable names for use during the
compare/merge
e.g.
table:map
original    rename
----------    ------------
X1    _X1
...
X10    _X10
foo    bar
hula    hoop
Y1    _Y1
...
Y7    _Y7

Then use sql to create macro variables to do the nitty gritty renaming and
reassigment

proc sql noprint;
  select trim(original)||'='||rename, trim(original)||'='||trim(rename)||';'
into :rename separated by ' ', :reassign separated by ' ' from map;

then in the data step
in the merge use YYYY(rename=(&rename))
and in the conditional assigment use &reassign;

Richard DeVenezia

"Sharavi Gandham" <SHARAVI@WORLDNET.ATT.NET> wrote in message
news:200005032206.SAA19940@listserv.cc.uga.edu...
> Hello All
>
> I have two datasets in this way
>
> XXXX                                 yyyy
> a 1                                 a 2
> b 1                                 b 2
> c 1 etc;.                           c 2 etc;.
>
> As you can see both the datasets have the smae variable names. I need to
> compare both the variables by keeping both values to see the
> cocordance/discordance between them.I know that i can merge them and
compare
> them.The problem is, i have 80 variables and i have to rename all of them
> for one of the dataset and then do a merge to compare.
> I can do that,but as you can see it is such a laborious process, I need to
> know if there is any easier way to do that.
>
> Thanks all in advance,
>
> Also i need to thank all the SAS listserv group people who have replied to
> me earlier too, i am glad to be a part of this group,Thanks again.