At 05:06 PM 7/28/2004, Kurt Wilkening wrote:

>I have one string variable called name that contains the last name >followed by a space, then first name. For example, > >Smith John R > >I would like to run syntax that creates two variables from the one so that >I end up with one variable called last name that contains "smith" and >another variable called first name that contains "John R".

I'm sure you'll have a correct answer posted by the time this hits the list, but I think it's worth a comment.

Extracting logically different parts from a string, based on punctuation or other context in the string, is called 'parsing', and it's fascinating, challenging, and cause good people to tear their hair out. I have *been* there.

Your problem, as you've stated it, is an easy one; here's the solution probably already posted. If your original variable is called FULLNAME, then:

. STRING LASTNAME FRSTNAME (A20). . NUMERIC #BRKPNT (F3). . COMPUTE #BRKPNT = INDEX (FULLNAME,' '). . COMPUTE LASTNAME = SUBSTR(FULLNAME,1,#BRKPNT). . COMPUTE FRSTNAME = SUBSTR(FULLNAME,#BRKPNT+1).

The only 'cute' trick is the scratch variable for the location of the first space.

>Can someone provide quick help?

OK, there's quick help. Now, here's non-quick help. Real inputs for parsing almost NEVER follow a simple rule consistently. The following just scratch the surface:

"Smith John R" -> "John R" "Smith" (correct)

* Parsing company and institution names can give ludicrous results(*):

"The Acme Co." -> "Acme Co." "The" "North American Van Lines" -> "American Van Lines" "North" "Supreme Court of Alabama" -> "Court of Alabama" "Supreme"

* A simple deviation: leading blanks on the name will give a blank last name,

" Smith John R" -> "John R Smith" " "

So,

. STRING LASTNAME FRSTNAME (A20). . STRING #TRMNAME (A50. . COMPUTE #TRMNAME=LTRIM(FULLNAME). . NUMERIC #BRKPNT (F3). . COMPUTE #BRKPNT = INDEX (#TRMNAME,' '). . COMPUTE LASTNAME = SUBSTR(#TRMNAME,1,#BRKPNT). . COMPUTE FRSTNAME = SUBSTR(#TRMNAME,#BRKPNT+1).

* Second simple deviation: TWO blanks between first and last name give a leading blank on first name,

"Smith John R" -> " John R" "Smith"

So, combining the previous correction with this one,

. STRING LASTNAME FRSTNAME (A20). . STRING #TRMNAME (A50. . COMPUTE #TRMNAME=LTRIM(FULLNAME). . NUMERIC #BRKPNT (F3). . COMPUTE #BRKPNT = INDEX (#TRMNAME,' '). . COMPUTE LASTNAME = SUBSTR(#TRMNAME,1,#BRKPNT). . COMPUTE FRSTNAME = LTRIM(SUBSTR(#TRMNAME,#BRKPNT+1)).

* Third deviation: A comma instead of blank separating the names:

"Smith,John R" -> "R" "Smith,John"

Adding this to the previous corrections,

. STRING LASTNAME FRSTNAME (A20). . STRING #TRMNAME (A50. . COMPUTE #TRMNAME=LTRIM(FULLNAME). . NUMERIC #BRKPNT (F3). . COMPUTE #BRKPNT = INDEX (#TRMNAME,' ,',2). . COMPUTE LASTNAME = SUBSTR(#TRMNAME,1,#BRKPNT). . COMPUTE FRSTNAME = LTRIM(SUBSTR(#TRMNAME,#BRKPNT+1)).

(Note the 'cute' use of the feature of INDEX.)

And that still doesn't fix

"Smith,John R" -> "R" "Smith,John" "Smith ,John R" -> ",John R" "Smith"

Good luck! If your input wasn't machine-generated, you may see every one of these, and almost certainly more that I haven't thought of. .......... (*) The John Carter Brown Library is an endowed rare-book library at Brown University. Years ago, it received a solicitation letter:

Mrs. John Carter Brown Library <correct address>

Dear Mrs. Library: