using sed to remove line in a comma-delimited file

seefor · 03-10-2009, 01:58 PM

Greetings all,
I have a script that output all my data in to a comma-delimited file separated by ";"

Current Output:

Quote:

"SAP 1117A";10.94.1.7;239.234.1.12;0;0;0;0;0;3;172.31.207.45;""
"";NA;239.192.1.50;0;0;0;0;0;3;172.31.207.61;""
"";NA;239.192.1.50;0;0;0;0;0;3;172.31.207.62;""
"";NA;239.192.1.50;0;0;0;0;0;3;172.31.207.63;""
"";NA;239.192.1.50;0;0;0;0;0;3;172.31.207.64;""
....

I would like to remove all lines that have an NA in the second field.
New Output

Quote:

"SAP 1117A";10.94.1.7;239.234.1.12;0;0;0;0;0;3;172.31.207.45;""
....

Would this be possible to do in sed?

Thanks in Advance,
SeeFor

jschiwal · 03-10-2009, 02:15 PM

I guess you would call this a semi-colon delimited file!

I think awk would be a better tool because the data is organized in records, and the NA needs to match in just the second field.

Code:

awk -F\; '$2 != "NA" { print }' sample
"SAP 1117A";10.94.1.7;239.234.1.12;0;0;0;0;0;3;172.31.207.45;""

crabboy · 03-10-2009, 02:25 PM

This could possibly match an NA not in the second field, but its quick.

Code:

| sed  '/^*;NA;*/d'

seefor · 03-10-2009, 02:42 PM

Thank you very much jschiwal and crabboy.

I couldn't use SED cause the rest of the data have NA in different fields.

Code:

awk -F\; '$2 != "NA" { print }'

Works great

Thanks again,
SeeFor

jschiwal · 03-10-2009, 03:35 PM

You could use a matching expression such as :sed '/^[^;][^;]*;NA;/d' file. But it would be uuugly.
Awk is the program to use for text files that are organized in fields.
That is what it was written for.