Shell Programming and Scripting

Hello ppl

I have a requirement to split (cut in unix) a file (A.txt) which is a pipe delimited file into A1.txt and A2.txt

Now I have to join (paste in unix) this A2.txt with external file A3.txt to form
output file A4.txt which should be CSV (comma separated file) so that third party can open it using Microsoft excel.

I can split the files however I want according to my flexibility but my input A.txt is a pipe delimited file and my output should be A4.txt.

the reason my input A.txt was pipe because the fields inside might contain commas within them .

Any quick help is appreciated..

Replace all pipes by commas

Code:

tr '|' ',' <input >output

In case the fields have commas try

Code:

sed 's/|/","/g; s/^/"/; s/$/"/'

please give samples files.

you wrote:
so output files can't have comma as separator in such case, unless you surrounds data with double-quote ( " ) for instance .

That's what the sed solution does, protectively quotes all fields.

You can try to add this filter that converts | to ","

Code:

... some code ...|awk -F\| -vOFS='","' '{print "\"" $1,$2,$3,$4 "\""}' > A4.txt

Jean-Paul

But, of course, the above script only works correctly if the input fed into it always has exactly four fields in the input (and I didn't see anything in this thread so far that places any limits on the number of fields).

The following should work (even keeping empty empty lines in the input empty in the output):

Code:

... some code ... | awk -F'|' -v OFS='","' 'NF{$1 = $1; $0 = "\"" $0 "\""}1' > A4.txt

If you want to try this on a Solaris/SunOS system, change awk to /usr/xpg4/bin/awk, /usr/xpg6/bin/awk, or nawk.

another thing I forgot to mention is I dont know pre hand how many fields exist in the file meaning A2.txt can have 0-n number of fields ..The file A2.txt might have any number of variable fields.