Go to ...

Techpository

A Better Technical Repository

RSS Feed

Linux: Using diff to compare files


Purpose:

diff reports the differences between two files.
Description

diff [options] <i>file1</i> <i>file2</i>

Simple Usage Example

File test1:

Test!
A common line.
Really common line.
One more common line.
Only here.

File test2:

Test!
A almost common line.
Really common line.
Really not common line.
One further line.
One more common line.

For example

$ diff test1 test2

will output:

2c2
< A common line.

> A almost common line.
3a4,5
> Really not common line.
> One further line.
5d6
< Only here.

What does that mean? First of all this output can be used as a script for ed (see example ). Lines
from first file are preceded by a less then symbol ( < ) and lines from
second file by a greater then symbol ( > ).
A dashed line ( — ) is used to separate output from the two files.
The letters can be used to convert file1 into file2:
c Replace lines from file1 with those from file2.
d Delete lines from file1.
a Add lines from file2 to file1.

The two files have three differences:

Line 2 differs.
Lines 4 and 5 of test2 are not present in test1. These lines would need to
be added after line 3 of test1 for the files to be the same.

Since you normally think in terms of converting the first file into the second file it is better to say that line 5 would need to be deleted from test1 for the files to be the same.

Useful Options

Option

Description

-b
Ignores repeated blanks (e.g. <space><space> is the same as <space>) and blanks at the end of lines.

-w
Ignore all spaces and tabs (e.g. 1 or 2 is equivalent to 1or2).

-i
Ignore case (e.g. howdy, HOWDY and HoWdY are equivalent).

-c
Use the context output format. Context output includes three lines before and after those that are normally printed to give “context” for the differences.

-u
Use the unified output format, easy readable with file information.

-C n
Like -c but include n lines of context output.

-e
Produce a script file that can be used by ed to convert file1 to file2. This option is not used as often as the patch command to convert file1 to file2.

-h
Do a faster but less accurate comparison. This does not work well if the files are very different and cannot be used with the -e option.

Directory related Options

Option

Description

-l
Output is formatted so that each file comparison occurs on a new page. Other comparisons are listed on a final page.

-r
Recursively compare all files in common subdirectories.

-s
Include a listing of all identical files in the output.

Examples

Comparing Directories

diff /tmp/oldFolder/ /tmp/newfolder/

Try it out! The output is self-explaining.

Changed files in folder tree

lists all files that have changed in a folder tree

diff -uwrq /tmp/oldFolder/ /tmp/newfolder/

Ignore Case and Repeated Blanks

Report the differences between poem1 and poem2 using the -i option to ignore the differences between upper and lower case charactersand the -b
option which ignores all repeated blanks and blanks at the end of lines.

diff -ib test1 test2

Patch with Diff

The recommended way …
Using output redirection

diff test1 test2 > diff_for_patch

patch test1 diff_for_patch

test1 is converted into test2 regarding the diff results, wich you can
influence by using options.

With

diff -r dir1 dir2 > dir2.patch

you can create a patch over all files in a directory which you can apply using

cd dir1 ; patch -p1 < dir2.patch

Ed with Diff

The -e option creates a script that gives directives to the ed text editor to convert file1 into file2.

diff -e test1 test2 > diff_for_ed

( cat diff_for_ed && echo w )| ed – test1

( ) for subshells, | – piping, ‘echo w’ is appended to ed input to make ed write the file

Diff with Context and more verbose Output

diff -c test1 test2

You can use the -C n option for n context lines. Output related to file1 is
preceded by stars (***) and file2 by dashes (—). Differences are separated by a long row of stars (***************).
In output the following symbols are used:

! Indicates corresponding lines in the two files that differ.
+ Indicates lines that exist in file2 but not file1.
– Indicates lines that exist in file1 but not file2.

Diff with Script

Using the Bourne shell:

#!/bin/sh
# use -h option for faster, less accurate comparison
diff -h $1 $2 &amp;gt; /dev/null
# Asking for exit status using $? for bourne shell, 0 means equal
if [ $? -eq 0 ]; then
echo identical
elif [ $? -eq 1 ]; then
echo different
else
echo an error occurred
fi

By: K Rekk