loop to replace repeating characters

[tradntele]

I have a file that has repeating commas and I'd like to replace them with a single comma (effectively, creating a csv file). The challenge is that there can be an infinite number of repeating commas. For example,

field1,field2
field1,,field2
field1,,,,,,,,,field2

So how can I do the substitution... I was thinking of using sed within a while loop, but that doesn't seem to work because of output redirection. For instance,

while grep ',,' datafile
do
sed 's/,,/,/g' datafile > updatedfile
done

If it loops, sed would never actually look at the updatedfile to continue the substitution and it would loop forever.

 

[vgersh99]

's/[,][,]*/,/g'
vlad
+---------------------------+
|#include<disclaimer.h> |
+---------------------------+

 

[dchoulette]

cat datafile | tr -s &quot;,&quot; &quot;,&quot; > updatedfile