Conversion to csv file

[CHARLYROSARIO]

I am working in a ksh script to get one an output csv file of ONLY one line like:
/,45,/usr,56,/var,41

The input file has the following information structure
/,45
/usr,56
/var,41

I was using the following command but it is showing all info overlapped and not separated by comma:

cat dftmp2 | awk '{ for (i=1; i <= NF; i ++) {if (i < NF) printf "%s,",$i;else printf "%s\n",$i}}'`

,/var,41

Any help

 

[PHV]

And what about this ?

awk '{x=x","$0}END{print substr(x,2)}' dftmp2

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886

 

[aigles]

A more cryptic solution using sed :

sed -n 'H;${x;s/\n//g;s/,$//;p;}'dftmp2

For a production script, prefer the PHV'solution that is more simple to understand and to maintain.

Jean-Pierre.

 

[PHV]

Jean-Pierre, in a legacy unix system the following works better:

sed -n 'H;${x;s/\n/,/g;s/^,//;s/,$//;p;}' dftmp2

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886

 

[Chapter11]

cat datafile | xargs echo | tr -d " "

 

[Chapter11]

Scrap that.

cat datafile | xargs echo | tr " " ","

 

[PHV]

A shorter variation Chapter11 (UUOC patrol again ;-)):
xargs < dftmp2 | tr ' ' ','

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886

 

[CHARLYROSARIO]

Thanks a lot to all solutions i have tested using a flat txt file and it worked it.
Now the new problem is the following:
As input file contains <CTRL> Characters at the end of the file, then the file manipulation with all these solutions canot be made:
65,/^M
18,/stand^M
0,/proc^M
0,/dev/fd^M
0,/dev/_tcp^M
0,/system/processor^M
36,/home^M

Is there any way to extract the ^M at the end of each file?
If anybody ask, the only reason why we have such ^M characters is that the info is being transferred by remote shell and unfortunately it is the only one to get it.

Thanks a lot in advance

 

[KenCunningham]

Does your system include dos2unix? This will strip the ^Ms from your file before processing.

I want to be good, is that not enough?

 

[hoinz]

Depending on your flavour of Unix, have a look at commands dos2unix or dos2ux.
(^M is needed for Dos, but not for Unix.)

 

[hoinz]

ok, Ken, you were faster ....

 

[KenCunningham]

...for a change!

I want to be good, is that not enough?

 

[CHARLYROSARIO]

Thanks a lot to Ken and hoinz for your tip. It worked great.
Thanks a lot to PHV, aigles and Chapter11 for your help on the csv format.
Best Regards

Charly

 

[Chapter11]

dos2unix is a crutch

stealing from PHV:

xargs < dftmp2 | tr ' ' ',' | tr -d '\r'

Or in general, pipe your source file through a

tr -d '\r'

to strip carriage returns.

 

[PHV]

A no pipe way:
awk '{x=x","$0}END{gsub(/\r/,"",x);print substr(x,2)}' dftmp2

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886