conversion

[krava]

Hi,

I would like to have bash function that would convert space-separated data table into comma-separated one. This is what I put together so far

function dat2csv() {
for fl in `find -type f -name "$1" -print`; do
#echo $fl;
bNm="$(basename $fl .dat)"; # echo $bNm;
of=$bNm".csv"; #echo $of;
echo $fl ">" $of;
sed 's/ /,/g' $fl > $of
done
}

This works but if I have table that is separated by 2 spaces than it fails. How to make this function useful in genaral case (separated by 1, 2 ... n spaces)?


thx
p.

 

[feherke]

Hi

So you want to transform like 1 space -> 1 comma, 2 space2 -> 1 comma, ..., n spaces -> 1 comma ? What happens if 2 spaces mean 2 separators and the data between them is empty string ? Anyway :

sed 's/ [red]\+[/red]/,/g' $fl > $of

[gray]# or[/gray]

sed 's/ [red][highlight #fee] [/highlight]*[/red]/,/g' $fl > $of

Feherke.

http://rootshell.be/~feherke/

 

[krava]

there is no empty string ... I am lucky I guess

this is what happens with sed 's/ \+/,/g' $fl > $of

119.77698 420.95663
4.93927 2.84465
24.99672 84.16383
.......... ........

converts to:

,119.77698,420.95663
,4.93927,2.84465
,24.99672,84.16383
...................

and with sed 's/ */,/g' $fl > $of to:

,1,1,9,.,7,7,6,9,8,4,2,0,.,9,5,6,6,3,
,4,.,9,3,9,2,7,2,.,8,4,4,6,5,
,2,4,.,9,9,6,7,2,8,4,.,1,6,3,8,3,
..............................

What should be done?

 

[PHV]

what about this ?
sed 's!^ \+!!;s! \+!,!g' $fl > $of

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

 

[p5wizard]

or with awk:

awk '{
for (i=1;i<NF;i++)
printf "%s,", $i
print $NF
}' /path/to/inputfile


HTH,

p5wizard

 

[feherke]

Hi

and with sed 's/ */,/g' $fl > $of to:

Maybe the pink background I used to emphasize the second space character was too light... You have to put two spaces in front of the asterisk ( * ).

Anyway, now you have versions which remove the leading separator, go with one of those.

Feherke.

http://rootshell.be/~feherke/

 

[krava]

thanks all suggs