deleting data and merging data

[omoo]

Hi,

I have the following data:

"header1";"header2";"header3";"header4";"data1";"data2";"1Lnum1";"num2";"num3";"num4";"1Lnum5";"num6";"num7";"num8"

I wan the final output to be:

"header1","header3","header4"
"1Lnum1.num2","num3","num4"
"1Lnum5.num6","num7","num8"

In summary, I want to merge all the data under header1 and header2 with a dot in between and delete all the unnecessary data (data1,data2) in between the last header and the first "1L data. I hope I am clear with my problem.

 

[p5wizard]

man awk

look for IFS and printf info



HTH,

p5wizard

 

[omoo]

Hi,

Can put some scripts so that I can work from there?

 

[p5wizard]

You should get the idea from this:

echo '"h1";"h2";"h3";"h4";"d1";"d2";"1L1";"n2";"n3";"n4";"1L5";"n6";"n7";"n8"' | awk -F';' '{
 printf "%s,%s,%s\n", $1, $3, $4
 printf "%s.%s,%s,%s\n", $7, $8, $9, $10
}'

start counting...

HTH,

p5wizard

 

[omoo]

Hi,

Thanks I have sort out the main part of the formatting. Now I only have one problem, merging the data:

input:
"num1","num2","num3"

output:
"num1num2","num3"

What I always end up getting is:
"num1""num2","num3"
but I do not want the "" in the middle of 1st field and 2nd field. Any advice?

 

[omoo]

just to clarify, the fields are not all with the " at both ends. That is to say for example:

input:
"num1","num2","num3",num4,num5,"num6"

 

[p5wizard]

# echo '"abcd"'
"abcd"

# echo '"abcd"'|awk '
{
printf "%s\n", substr($1,2,length($1)-2)
}'
abcd

You take it from here...

HTH,

p5wizard

 

[dchoulette]

You could try:

sed 's/^"//;s/"$//' | awk -F '"?,"?'

in the first example given by p5wizard.

The sed suppress any " at start or end of line and the -F in awk allowed for optional " before and after the comma.

This simple trick would not work if there is any comma or escaped quotes inside the quoted fields of your csv file.
For example:
[tt]"value 1","value 2,3 & 4","He said \"hello\""[/tt]
And you're in trouble.

--------------------

Denis