["Followup-To:" header set to comp.unix.shell.]
On 2008-03-29, sandy_eggo <cpmadrid@[EMAIL PROTECTED]
> wrote:
>
>
> [cross-post comp.lang.awk, comp.unix.shell]
>
> I have a file whose record delimiters seem to be "Directions" and
> "More" (i.e., I need the data between each, however they need to be
> joined and comma-delimited. Kind of the inverse of csv parser.
>
> For example, given:
>
> More Info | Map | Directions
>
> Milford Pharmacy & Wellness Center
> 931 State Rte 28 Ste 205
> Milford, OH 45150-1990
> (513) 831-8211
>
> More Info | Map | Directions
>
> Alexandria Discount Drugs
> 7857 Alexandria Pike
> Alxndra, KY 41001-1045
> (859) 635-2171
>
> More Info | Map | Directions
>
> should return
>
> "Milford Pharmacy & Wellness Center","931 State Rte 28 Ste
> 205","Milford","OH",45150-1990","(513) 831-8211"
> "Alexandria Discount Drugs","7857 Alexandria
> Pike","Alxndra","KY","41001-1045","(859) 635-2171"
>
> Note--There are quotes around the city, state, and zip to create
> discreet fields.
You could use a blank line as the record separator and discard records
beginning with "More Info".
awk -F\\n 'BEGIN{RS="";OFS="\",\""}
/^More Info/{next}
{$0="\"" $0 "\""; sub(/,/,"",$3);gsub(/ /,OFS,$3)}
{print}'
|
