CODEWITHSUNDEEP
bashshellfilehex
Ablia
Ablia

Reputation: 305

How to write an hex file

I have a file which is not "readable" (I mean, you can't open it with a text editor, it's weirdly encoded) but with hexdump, you can see some readable things. In this file, there is a date. Let's say "05082019". So when i do a hexdump -C i see this date, and the hex code associated: 3035303832303139

My goal is to write another file exactly like the first one, but with another date in it.

I've tried a simple sed on that date, like:

sed 's/05082019/04102022/g' myfile > newfile

But, since the file is "weirdly encoded", it seems sed can't find the date in it. So i've tried with hexdump:

hexdump -ve '1/1 "%.2x"' myfile | sed 's/3035303832303139/3034313032303232/g' > newfile

But that was a bit stupid: it's writing the hex string in newfile, giving it a totally different value.

i've also tried to change directly the file, like that: 

cp myfile newfile
 hexdump -ve '1/1 "%.2x"' newfile | sed -i 's/3035303832303139/3034313032303232/g'

But here sed -i is waiting for an input file.

Is there a way to write a file knowing only his hex value? (Or do you have another idea to solve that problem?)

(Note that this should be done in bash script. I'm sorry i can't give any encoded file as example.)

Upvotes: 0

Views: 450

Answers (2)

tshiono
tshiono

Reputation: 22012

You are almost there. Please try xxd -p to convert binary file to ascii representation and xxd -p -r to do the reverse.

xxd -p myfile | tr -d $'\n' | sed 's/3035303832303139/3034313032303232/g' | xxd -p -r > newfile

will do what you want.
Note that if your file size exceeds line length limitation of sed (depends on the sed version) the approach above may not work. Then please try a perl solution as an alternative:

perl -0777 -pe 's/05082019/04102022/' myfile > newfile

Unlike sed, perl can natively deal with binary files.

Upvotes: 1

Ross Jacobs
Ross Jacobs

Reputation: 3186

Replicating the File

If you use hexdump -C $file and get 3035303832303139, then $file MUST contain the plaintext 05082019 with no "special encoding".

We can replicate this file like so:

$ printf "Here are some random words 05082019 around the date" > myfile

Replacing Date With sed

You were on the right track, you just added an extra > newfile. Replacing it inline is then a matter of using sed -i.

$ sed -i -e 's/05082019/04102022/g' myfile

Note: -e used for compatibility between Macos/GNU sed.

Verify

$ cat myfile
Here are some random words 04102022 around the date

Upvotes: 2

Related Questions