Bash就地替换文本

Question

I have some insert statements which I want to transform for another environment by replacing values in place. My sql insert statement looks like below:-

INSERT INTO TABLE (col1, col2, col3, col4, col5) VALUES ('1', '/var/tmp', 'xyz, mno, pqr', '123', '<dummy value>');

I want to convert it such that there are no spaces left between column and values between () which would look like this:-

INSERT INTO TABLE (col1,col2,col3,col4,col5) VALUES ('1','/var/tmp','xyz,mno,pqr','123','<dummy value>');

What I have tried so far is something like this:-

echo "INSERT INTO TABLE (col1, col2, col3, col4, col5) VALUES ('1', '/var/tmp', 'xyz, mno, pqr', '123', '<dummy value>')" | awk -F'[()]' '{printf "%s(%s)%s(%s)", $1, $2, $3, gensub(".*\(","",gensub(" ", "", $4))}'

And the resulting output is as below:-

INSERT INTO TABLE (col1, col2, col3, col4, col5) VALUES ('1', '/var/tmp', 'xyz, mno, pqr', '123', '<dummy value>'))

I am stuck here. Any cues would be helpful.

Answer 1

I think you can have much simpler solution using sed :

sed -E 's/, /,/g'

If you're not sure about the comma's position, you can use a regex (see @anubhava comment ):

sed -E 's/ *, */,/g'

Example:

$ echo "(1, 2, 3) and (4 ,5 ,6)" | sed -E 's/ *, */,/g'
(1,2,3) and (4,5,6)