UNIX for Dummies Questions & Answers

I am using the following SED cript to extract sequences from a file:

Code:

sed -n '/string/p; //{n;p}'

For AWK, I expected this to work:

Code:

awk '!/string/{getline;next}1'

The file is something like this:

Code:

>string
AGAAGAGTGA
>stuff1
CGCGCG
>Stuff 2
AAAAAA

Expected output:

Code:

>string
AGAAGAGTGA

My SED script is producing the expected output but not AWK

Code:

>string

What I am missing? Is there a better way to do this? I will be screening hundreds of thousands of files and I would like to maximize the time as much as possible
Thanks!

Making a few wild guesses on incompletely specified requirements, you might want to try:

Code:

awk '/^>string/{p=2}p-->0'

The awk is not working because it is skipping all lines except the ones that contain "string", so for example when it reads

Code:

AGAAGAGTGA

it will skip that line...

I suspect that you were looking for something like this:

Code:

awk '/string/{print; getline; print}' file

--
However, since the payload of the records in your file may consist of multiple lines according to the format specification, perhaps it is better to try this instead and use ">" as record separator..

Code:

awk '/string/{print RS $0}' RS=\> ORS= file

or theoretically better:

Code:

awk '$1~/string/{print RS $0}' FS='\n' RS=\> ORS= file

--
Incidentally, the sed statement could be further reduced to:

Code:

sed -n '/string/{N;p;}' file

Thanks!