hexdump examples

The hexdump man page is a bit tricky to follow as a way to get started with the command, but some better help can be found.

One tip for reading the man page is that the first few EXAMPLES require the -f option.

Here are a few more examples to jump-start new users.

The input file (hexdump.txt)

The correction for the aberration of light is said,
on high authority, not to be perfect even in that most perfect organ, the
eye.

Decimal byte offsets, 16 bytes per line

  1. Take 1 item of 0 bytes, print the current decimal byte offset padded to 10 digits with 0's, and then print the | pipe character. Next, take 16 items, each 1 byte in size, and print each item as a printing character. Finally, take 1 item of 0 bytes, and print another | pipe character followed by a newline.
hexdump -v -e '1/0 "%010_ad  |" 16/1 "%_p" 1/0 "|\n"' hexdump.txt
0000000000  |The correction f|
0000000016  |or the aberratio|
0000000032  |n of light is sa|
0000000048  |id,.on high auth|
0000000064  |ority, not to be|
0000000080  | perfect even in|
0000000096  | that most perfe|
0000000112  |ct organ, the.ey|
0000000128  |e..|

The same thing, omitting iteration counts of 1

The iteration count for each format unit (term from the man page) defaults to 1 if none is specified. So the iteration count can be omitted from both the first and last format units.

  1. Take 0 bytes, print the current decimal byte offset padded to 10 digits with 0's, and then print the | pipe character. Next, take 16 items, each 1 byte in size, and print each item as a printing character. Finally, take 0 bytes, and print another | pipe character followed by a newline.
hexdump -v -e '/0 "%010_ad  |" 16/1 "%_p" /0 "|\n"' hexdump.txt
0000000000  |The correction f|
0000000016  |or the aberratio|
0000000032  |n of light is sa|
0000000048  |id,.on high auth|
0000000064  |ority, not to be|
0000000080  | perfect even in|
0000000096  | that most perfe|
0000000112  |ct organ, the.ey|
0000000128  |e..|

The same thing again, now also omitting byte counts of 0

The byte count for each format unit defaults to 0 if the format unit contains no conversion characters other than %_a. So the byte count can also be omitted from both the first and last format units.

  1. Print the current decimal byte offset padded to 10 digits with 0's, and then print the | pipe character. Next, take 16 items, each 1 byte in size, and print each item as a printing character. Finally, print another | pipe character followed by a newline.
hexdump -v -e '"%010_ad  |" 16/1 "%_p" "|\n"' hexdump.txt
0000000000  |The correction f|
0000000016  |or the aberratio|
0000000032  |n of light is sa|
0000000048  |id,.on high auth|
0000000064  |ority, not to be|
0000000080  | perfect even in|
0000000096  | that most perfe|
0000000112  |ct organ, the.ey|
0000000128  |e..|

The same thing, now broken into two expressions

  1. Take 1 item of 16 bytes and print the current decimal byte offset padded to 10 digits with 0's, and then print the | pipe character.
  2. Take those same 16 bytes again, but this time divide them into 1-byte items, and print each item as a printing character. Then take 1 item of 0 bytes, and print a | pipe character followed by a newline.

Note that the starting byte offset only changes after both expressions have been evaluated, so the 16 printing characters still start at offset 0.

hexdump -v -e '/16 "%010_ad  |"' -e '16/1 "%_p" "|\n"' hexdump.txt
0000000000  |The correction f|
0000000016  |or the aberratio|
0000000032  |n of light is sa|
0000000048  |id,.on high auth|
0000000064  |ority, not to be|
0000000080  | perfect even in|
0000000096  | that most perfe|
0000000112  |ct organ, the.ey|
0000000128  |e..|

The same thing yet again, broken into three expressions

  1. Take 1 item of 16 bytes and print the current decimal byte offset padded to 10 digits with 0's, and then print the | pipe character.
  2. Take those same 16 bytes again, but this time divide them into 1-byte items, and print each item as a printing character.
  3. Take the same 16 bytes once more, go back to counting them as just 1 item of 16 bytes, and print a | pipe character followed by a newline.
hexdump -v -e '/16 "%010_ad  |"' -e '16/1 "%_p"' -e '/16 "|\n"' hexdump.txt
0000000000  |The correction f|
0000000016  |or the aberratio|
0000000032  |n of light is sa|
0000000048  |id,.on high auth|
0000000064  |ority, not to be|
0000000080  | perfect even in|
0000000096  | that most perfe|
0000000112  |ct organ, the.ey|
0000000128  |e..|

Again as three expressions, but using only one iteration/byte count

hexdump -v -e '/16 "%010_ad  |"' -e '"%_p"' -e '"|\n"' hexdump.txt
or 
hexdump -v -e '"%010_ad  |"' -e '16/1 "%_p"' -e '"|\n"' hexdump.txt
or 
hexdump -v -e '"%010_ad  |"' -e '"%_p"' -e '/16 "|\n"' hexdump.txt

All three of these are identical because of the way hexdump decides the size of the block (term from the man page) to use when reading through the file. hexdump takes a block of bytes matching size needed for the largest expression and passes it to each expression before it processes another block.

Offset printed for each byte, left-aligned characters, 7 bytes per line

hexdump -v -e '7/1 "%5_ad:%-5_c" "\n"' hexdump.txt
    0:T        1:h        2:e        3:         4:c        5:o        6:r    
    7:r        8:e        9:c       10:t       11:i       12:o       13:n    
   14:        15:f       16:o       17:r       18:        19:t       20:h    
   21:e       22:        23:a       24:b       25:e       26:r       27:r    
   28:a       29:t       30:i       31:o       32:n       33:        34:o    
   35:f       36:        37:l       38:i       39:g       40:h       41:t    
   42:        43:i       44:s       45:        46:s       47:a       48:i    
   49:d       50:,       51:\n      52:o       53:n       54:        55:h    
   56:i       57:g       58:h       59:        60:a       61:u       62:t    
   63:h       64:o       65:r       66:i       67:t       68:y       69:,    
   70:        71:n       72:o       73:t       74:        75:t       76:o    
   77:        78:b       79:e       80:        81:p       82:e       83:r    
   84:f       85:e       86:c       87:t       88:        89:e       90:v    
   91:e       92:n       93:        94:i       95:n       96:        97:t    
   98:h       99:a      100:t      101:       102:m      103:o      104:s    
  105:t      106:       107:p      108:e      109:r      110:f      111:e    
  112:c      113:t      114:       115:o      116:r      117:g      118:a    
  119:n      120:,      121:       122:t      123:h      124:e      125:\n   
  126:e      127:y      128:e      129:.      130:\n        :          :

The effect of -v

With -v

( echo "abcdefghijklmno"; echo "abcdefghijklmno" ) | hexdump -v -C
00000000  61 62 63 64 65 66 67 68  69 6a 6b 6c 6d 6e 6f 0a  |abcdefghijklmno.|
00000010  61 62 63 64 65 66 67 68  69 6a 6b 6c 6d 6e 6f 0a  |abcdefghijklmno.|
00000020

Without -v

( echo "abcdefghijklmno"; echo "abcdefghijklmno" ) | hexdump -C
00000000  61 62 63 64 65 66 67 68  69 6a 6b 6c 6d 6e 6f 0a  |abcdefghijklmno.|
*
00000020

One hexdump vs. od example

hexdump  -e ' "%07_ad" 16/1 "  %2_c" "\n"' hexdump.txt
0000000   T   h   e       c   o   r   r   e   c   t   i   o   n       f
0000016   o   r       t   h   e       a   b   e   r   r   a   t   i   o
0000032   n       o   f       l   i   g   h   t       i   s       s   a
0000048   i   d   ,  \n   o   n       h   i   g   h       a   u   t   h
0000064   o   r   i   t   y   ,       n   o   t       t   o       b   e
0000080       p   e   r   f   e   c   t       e   v   e   n       i   n
0000096       t   h   a   t       m   o   s   t       p   e   r   f   e
0000112   c   t       o   r   g   a   n   ,       t   h   e  \n   e   y
0000128   e   .  \n
od -Ad -w16 -tc hexdump.txt
0000000   T   h   e       c   o   r   r   e   c   t   i   o   n       f
0000016   o   r       t   h   e       a   b   e   r   r   a   t   i   o
0000032   n       o   f       l   i   g   h   t       i   s       s   a
0000048   i   d   ,  \n   o   n       h   i   g   h       a   u   t   h
0000064   o   r   i   t   y   ,       n   o   t       t   o       b   e
0000080       p   e   r   f   e   c   t       e   v   e   n       i   n
0000096       t   h   a   t       m   o   s   t       p   e   r   f   e
0000112   c   t       o   r   g   a   n   ,       t   h   e  \n   e   y
0000128   e   .  \n
0000131

Links

Manual on How to Use the Hexdump Unix Utility
More hexdump explanation and examples.
Debian hexdump man page
The Debian version of the hexdump man page has a few additions and formatting improvements compared to the FreeBSD version.
FreeBSD hexdump man page
The upstream version of the hexdump man page. Note that the +o strings on this page are just list bullets, not conversion characters.
od invocation
The od documentation from the GNU coreutils manual.
od man page
The Debian od man page.

Found a mistake?

Submit a comment or correction

Updates

2019 Sep 28 Correct byes to bytes thanks to a visitor comment. Rearrange the additional examples to show them in the one-expression format.
2019 Mar 09 Add more ways to write the two-expression and three-expression examples. Update the external links.
2013 Aug 17 Explanation rewording for the three-expression example
2013 Jan 08 Comments link
2012 Feb 04 more explanation, block size examples, od example, links section
2010 Dec 01 single-expression example and some rewording
2010 Nov 13 Fixed more missing \n escapes, hopefully fixed the root problem
2010 Oct 06 Fixed missing \n newline escapes
2010 Sep 13 Initial post