Here is a quick pipe that picks off by length of the word using NLOCATE
/*               */
aa.1 = ''          
aa.2 = '1'         
aa.3 = '21'        
aa.4 = '321'       
aa.5 = '4321'      
aa.6 = '54321'     
aa.0 = '6'         
'PIPE (End ?)    ',
'     Stem aa.   ',
'| n1:Nlocate 1  ',
'|    Spec \0\ 1 ',
'| f: Faninany   ',
'|    Console    ',
'? n1:           ',
'| n2:Nlocate 2  ',
'|    Spec \1\ 1 ',
'          1-* nw',
'|    f:         ',
'? n2:           ',
'| n3:Nlocate 3  ',
'|    Spec \2\ 1 ',
'          1-* nw',
'|    f:         ',
'? n3:           ',
'| n4:Nlocate 4  ',
'|    Spec \3\ 1 ',
'          1-* nw',
'|    f:         ',
'? n4:           ',
'|    Spec \>3\ 1',
'          1-* nw',
'|    f:         ',
''                 

Rollin Hill
----- Original Message ----- 
  From: Russel Brooks 
  To: [email protected] 
  Sent: Friday, April 03, 2009 2:02 PM
  Subject: Select records on word length?


  I am currently splitting records into separate words and stripping leading
  and trailing blanks and then dumping them into a stem where a Rexx loop
  looks at the word length.
  Is there a way I can filter on word length in a pipe stage?

  PIPE > fn ft fm | split | strip | <filter on length?> | ...

  --
  Russ Brooks  --  SVM Support  (aka: RLB at STLVM5)
  408-390-8483
  [email protected]

  Astronomy Picture of the Day: http://antwrp.gsfc.nasa.gov/apod/
  TANSTAAFL

Reply via email to