On Tue, Apr 15, 2008 at 1:09 PM, Mark J. Reed <[log in to unmask]> wrote:
> If I was right about the goal, here's one way to do it from the command line:
>
>  $ cat /file/with/just/numbers /file/with/pairs | awk '(NF==1)
>  {seen[$1]=1} (NF==2 && seen[$1]) {print}'

or just

awk '(NF==1){seen[$1]=1}(NF==2 && seen[$1]){print}' /file/with/numbers
/file/with/pairs >/output/file

which saves a pointless cat process.  That puts the result in
/output/file; change as desired, or lop off entirely to just spit out
the matching lines.  Here's a snippet of (untested) AppleScript that
should yield an array of records containing 'number' and 'letters'
fields:

set matchingLines to paragraphs of do shell script "awk
'(NF==1){seen[$1]=1}(NF==2 && seen[$1]){print}' " & quoted form of
POSIX path of fileWithNumbers & " " & quoted form of POSIX path of
fileWithPairs
set matchingPairs to {}
set oldTID to AppleScript's text item delimiters
set AppleScript's text item delimiters to space
repeat with aLine in matchingLines
    set end of matchingPairs to {number:  text item 1 of aLine,
letters: text item 2 of aLine}
end repeat
set AppleScript's text item delimiters to oldTID

-- 
Mark J. Reed <[log in to unmask]>