comparing of observation in single SAS data file

[Vijendra_SAS]

usually we use Proc Compare for comparing two sas datasets and their observation, variable.

is there any way we can compare the observation with in single sas datasets with eachother ? which procedure we use ? what is syntax ?

please let me know ASAP

[Reeza]

That depends on what you're trying to do.

What are you trying to do? Compare one variable against another variable? Compare one row against a different/all rows?

[Vijendra_SAS]

hi

I am comparing one row against another row..

[Reeza]

There are ways to make proc compare do that, but you'll need to replicate your dataset and indicate which rows are compared against which.

It's probably easier to go with a manual data step or a sql step.

If you want further advice, you'll need to provide more detailed information.

[Vijendra_SAS]

i have file , I couldnot attach it to conversation.. let me know your mail id.. I can send it across.. and we can discuss

[Reeza]

No thanks, attach here (edit your first post), or preferably recreate a simple sample that you can paste into the window.

IE usually won't allow you to paste, so you might need to use FireFox or Chrome.

[Vijendra_SAS]

Hi

i tried to attach file in my first post, you may refer there , about what I want to know.

[Vijendra_SAS]

In attached file , I want to check if Field_Name across all the Form_Name have same definition or not ? if its varies even by space, comma or fullstop , my SAS program should give me out file with those difference.

I have applied filter , just for convenience of your understanding.

let me know which function I can use ?  ( I am aware of lag function.. but not able to put inform of sytax )

[Reeza]

You don't specify what you want your output to look like.

Sort your data by Field_Name and use first./last. processing in addition to the definition. If the definition varies while the field name is the same it will flag it.

I don't think this is quite right, but you can review the first/last processing and see what works for you.

proc sort data=have; by form_name field_name definition; run;

data want;

set have;

by form_name field_name;

if first.field_name and not (first.definition and last.definition) then flag=1;

else flag=0;

run;

[Astounding]

Maybe what you really want is as simple as this:

proc freq data=have;

   tables field_name;

run;

If not, it might be this:

proc freq data=have;

   tables form_name * field_name / missing list;

run;

It won't point out which rows you need to fix, but it will find the spelling variations.