split a column into three columns

[France]

Dear all,

I would like to split the following column (i.e., table1) into three columns(i.e., table2), 

table 1
NAME
DYNAMIC COML.FIN. DEAD - DELIST 23/09/04
THALES (LON) DEAD - DEAD 28/07/03
EDIRECTORY CO UK DEAD - 20/05/08
INTERNET TIMES DEAD - 29/08/03
CYBERES DEAD - DELIST 24/12/04
MTR CORP. (XSQ) DEAD - 01/10/08
INTERCEDE GROUP
XPLOITE DEAD - T/O BY 31135W
TIGER RESOURCE
GALAHAD GOLD DEAD - 21/02/08
INNOBOX DEAD - 19/05/08
PORTMAN (LON) DEAD - DELIST 02/08/04
WARTHOG DEAD - 13/05/08
LASMO ASSD AGIP INV.CSH
MCBRIDE
SUMITOMO CORP. (UK) LTD
NATIONAL GRID (WI)
IST MOBL.ITAL.ADS (LON) DEAD - DELIST 06/02/98
IST MOBL.ITAL.SPA (LON) DEAD - DELIST 03/07/98
ITALGAS (LON) DEAD - DELIST 03/07/98
MITSUBISHI ESTATES (XSQ) DEAD - 01/10/08
MITSUBISHI HEAVY (XSQ) DEAD - 01/10/08
MITSUBISHI MOTORS (XSQ) DEAD - 01/10/08
MITSUI & COMPANY (XSQ) DEAD - 01/10/08

like

table 2
NAME, STATUS, DATE
DYNAMIC COML.FIN., DEAD - DELIST, 23/09/04
THALES (LON) DEAD, DEAD, 28/07/03
EDIRECTORY CO UK, DEAD,  20/05/08
INTERNET TIMES, DEAD, 29/08/03
CYBERES, DEAD - DELIST, 24/12/04
MTR CORP. (XSQ), DEAD, 01/10/08
INTERCEDE GROUP
XPLOITE DEAD - T/O BY 31135W
TIGER RESOURCE
GALAHAD GOLD, DEAD, 21/02/08
INNOBOX, DEAD, 19/05/08

Thanks in advance.

[PaigeMiller]

What are the rules used to do this?

[RW9]

Post test data in the form of a datastep in future please.  I cannot see any logical way to split that data as there is no consistency.  I would advise you go back to the source and fix, or get fixed, there.

[Shmuel]

Your three variables are:  Name, Status, Date.

You probably know what statuses are valid - according to your sample any status starts with 'DEAD'.

Date always start with a number.

Assuming no name include status string and the order is always Name then Status then Date

you can try next code:

data want;
  infile datalines truncover;
  input a_line $50;
  
  p1 = index(a_line, 'DEAD');         /* status start position */
  p2 = indexc(a_line, '0123456789');  /* date start position */
  
  Name = substr(a_line,1,p1-1);
  Status = substr(a_line,p1,p2-p1);
  Date = input(strip(substr(a_line,p2)),ddmmyy8.);
  format date ddmmyy10.;
  
datalines;
DYNAMIC COML.FIN. DEAD - DELIST 23/09/04
THALES (LON) DEAD - DEAD 28/07/03
EDIRECTORY CO UK DEAD - 20/05/08
INTERNET TIMES DEAD - 29/08/03
CYBERES DEAD - DELIST 24/12/04
MTR CORP. (XSQ) DEAD - 01/10/08
INTERCEDE GROUP
XPLOITE DEAD - T/O BY 31135W
TIGER RESOURCE
GALAHAD GOLD DEAD - 21/02/08
INNOBOX DEAD - 19/05/08
PORTMAN (LON) DEAD - DELIST 02/08/04
WARTHOG DEAD - 13/05/08
LASMO ASSD AGIP INV.CSH
MCBRIDE
SUMITOMO CORP. (UK) LTD
NATIONAL GRID (WI)
IST MOBL.ITAL.ADS (LON) DEAD - DELIST 06/02/98
IST MOBL.ITAL.SPA (LON) DEAD - DELIST 03/07/98
ITALGAS (LON) DEAD - DELIST 03/07/98
MITSUBISHI ESTATES (XSQ) DEAD - 01/10/08
MITSUBISHI HEAVY (XSQ) DEAD - 01/10/08
MITSUBISHI MOTORS (XSQ) DEAD - 01/10/08
MITSUI & COMPANY (XSQ) DEAD - 01/10/08
;
run;

[Shmuel]

I see there are some exceptions so the code should be adapted to check is p1=0 (no status) then status should be blank and name=a_line.
Is it possible to have a date without status ?

[Shmuel]

Fixed code is:

data want;
  keep Name Status Date;
  length Name $20 status $15;  /* adapt to max expected */

  infile datalines truncover;
  input a_line $50;
  
  p1 = index(a_line, 'DEAD');         /* status start position */
  if p1 = 0 then Name = trim(a_line);
  else  Name = substr(a_line,1,p1-1);
  
  p2 = indexc(a_line, '0123');  /* date start position, day is 0n/1n/2n/3n */
  if p1 = 0 or p2=0 then date=.;
  else do:
       Status = substr(a_line,p1,p2-p1);
       Date = input(strip(substr(a_line,p2)),ddmmyy8.);
    end;
  format date ddmmyy10.;
  
datalines;

...........

; run;