[SOLVED] substring using awk and sed ( binary )

iffarrukh · 05-14-2012, 12:32 AM

Hi..
I got a problem and I am looking some help. I was using this command to remove @ from binary file and then getting substr and finally using printf formating it and putting into a file. I am trying to run this using new version of awk(3.1.7) and I am just getting 4 character while previously I was almost 8727( awk 3.1.3)
I know awk is not supposed to work with binaries but code is written year ago and I cannot change a lot..

$ cat TT120T0-inbound | sed -e "s/@@//g;s/@$//" | awk ' begin {} {x = substr($0,1,length($0) - 4); printf("%s", x); }' >> Final-data

Any help..

I have attached the data as well but its a txt file not binary file so might not work properly . but just for reference. Have a look at snapshot as well.

Thanks in advance

Regards
Sam

millgates · 05-14-2012, 03:47 AM

Quote:

Originally Posted by iffarrukh

Hi..
I got a problem and I am looking some help. I was using this command to remove @ from binary file and then getting substr and finally using printf formating it and putting into a file. I am trying to run this using new version of awk(3.1.7) and I am just getting 4 character while previously I was almost 8727( awk 3.1.3)
I know awk is not supposed to work with binaries but code is written year ago and I cannot change a lot..

$ cat TT120T0-inbound | sed -e "s/@@//g;s/@$//" | awk ' begin {} {x = substr($0,1,length($0) - 4); printf("%s", x); }' >> Final-data

There is no need to use cat, awk and sed. The same can be done with a single awk or sed command:

Code:

sed -e "s/@@//g;s/@$//;s/.\{4\}$//" TT120T0-inbound # this does not remove the newlines, though
                                                    # I am not sure if that is what you want
# or

awk '{ gsub(/@@/,"",$0); sub(/@$/, "", $0); sub(/....$/,"",$0); printf $0 }' TT120T0-inbound

Also, the lines in your file are quite long. Some implementations of awk or sed might have a problem with that.

Quote:

Originally Posted by iffarrukh

Hi..
I have attached the data as well but its a txt file not binary file so might not work properly . but just for reference. Have a look at snapshot as well.

Ok, then what is the difference between the binary and text versions?

pan64 · 05-14-2012, 04:47 AM

and also wc -c <filename> works. you do not need cat <filename> | to use wc

iffarrukh · 05-14-2012, 11:11 PM

Hi millgates

Thanks for your help.. Sed command is working as I was looking for..

Would you mind telling me how this is working as I am having a bit confusion abt it..

sed -e "s/@@//g;s/@$//;s/.\{4\}$//" TT120T0-inbound

Regards
Sam

pan64 · 05-15-2012, 12:28 AM

sed has 3 commands:

Code:

s/@@//g
s/@$//
s/.\{4\}$//

the first one substitutes @@ with nothing, so it removes @@ from everywhere.
the second will remove the @ from the end of the lines
this two was taken from your original post

the third removes 4 characters from the end of the line.

iffarrukh · 05-15-2012, 06:15 PM

Hi

Thanks all for your support and help

I was just curious about third part of sed.

Cheers
Sam

sycamorex · 05-15-2012, 06:25 PM

I think you could combine the first 2 sed commands:

Code:

sed -r 's/(@@|@$)//g' file

pan64 · 05-15-2012, 11:22 PM

Happy with solution ... mark as SOLVED

If someone helps you, or you approve of what's posted, click the "Add to Reputation" button, on the left of the post.

I'm glad to help you