Shell Programming and Scripting

Hi Friends,

I am using a command that prints certain lines from a file.

For ex:

cat input

Code:

abc chr1 456
def chr1 789
ghi chr1 999
jjj chr1 777
jhk chr7 914

My command

Code:

awk '{if($2=="chr1" && $3>=456 && $3<=999) {print $0}}' OFS="\t" input

Output being printed is

Code:

abc chr1 456
def chr1 789
ghi chr1 999

I want to pipe this output to sed or whatever to delete these lines. Maybe like this

Code:

awk '{if($2=="chr1" && $3>=456 && $3<=999) {print $0}}' OFS="\t" input | sed '/d/' 

My final required input file should be

Code:

jjj chr1 777
jhk chr7 914

I don't want to try it for myself because it is a huge 27GB input file. Any mistakes will cost me money and time to regenerate it.

I'd strongly suggest you try your OWN implementation on a smaller sample and work out the 'kinks' yourself - it should be pretty simple given you already have a good start that you presumably understand.
Not trying it yourself will cost you knowledge in the long run.
Good luck and do come back if/when you get stuck.

Your awk logic doesnt make any sense to me...

your requirement is going nowhere...or am not understanding...

The OP needs to reverse the original awk logic - the implementation is left as an exercise.

Hi vgersh99,

But, how do I feed the awk's output to sed? It is giving me the following error when I pipe to

Code:

sed '/d'

Code:

sed: -e expression #1, char 3: missing command

and then I used -e and still it gives the same error.

---------- Post updated at 03:36 PM ---------- Previous update was at 03:24 PM ----------

Okay. Thanks for the suggestions.

I now understood what needs to be done.

So, what I did was this

Code:

awk '{if($2=="chr1" && $3>=456 && $3<=999) {print $0}}' OFS="\t" input > intermediate_input

Code:

join -v1 -v2 input intermediate_input > final_input

But, there is a small glitch here. I am missing the immediate next line after the last line in intermediate_input.

Which means, instead of

Code:

jjj chr1 777
jhk chr7 914

I am seeing

Code:

jhk chr7 914

Any thoughts?

Alos, please suggest a pointer to keep the input file and the final_input file's pattern or separators to be the same. They changed completely after running the above command.

you don't need anything other than awk:

Code:

awk '!($2=="chr1" && $3>=456 && $3<=999)' OFS='\t' myFile

In your desired output"

Code:

jjj chr1 777
jhk chr7 914

why do you need the line in red? This is not what you coded...