How to understand absolute and relative paths in Linux
Absolute and Relative Paths
In Linux there are two ways to address a file, (1) using a relative path, and (2) using absolute path.
If you have a file in the subdirectory my_data and the file name is my_test.dat.
Then you can use the absolute path
/home/user.name/my_data/my_test.dat
This path is absolute, as it starts at the very top, /, the root directory.
Users on ARC do not have permissions to write to /,
but they do have permissions to write to their own /home/user.name/ home directory,
and everything below it.
While it is very precise and clear when absolute paths are used, they tend to be long and also they really hard-wire the locations of the files to the specific directory and the directory cannot be moved easily to a different storage location without changing any references to it.
The relative paths can be really a convenient alternative, therefore.
They are always relative to some other path.
For example,
if a user just logged into his/her account the bash command line starts in the home directory,
which is /home/user.name, therefore, the relative path to that file will be
my_data/my_test.dat
if we change the command line to the directory my_data, then the relative path becomes simply
my_test.dat
Please note, that any path that starts with / is an absolute path, by definition.
When using the command line, one can always check the current working directory of the session with the
$ pwd
command, that stands for Print Working Directory.
This way one can see the reference location for the relative file and directory names.
Special Relative Paths
There are two very commonly used relative paths that have to be mentioned, ., the dot path, means this directory,
and the .., the double dot path, which means the parent directory, or the directory above.
Thus, the ./my_test.dat is fully equivalent to simply my_test.dat.
Special Absolute Paths
The "~" (the tilda character) is used to point to the home directory. It is fully equivalent to the /home/name.user/ path. The file from the example can also be referenced as
~/my_data/my_test.dat
rsync and scp Remote Paths
When we transfer files using rsync or scp commands,
we provide the account details as user.name@arc.ucalgary.ca,
which means, the user user.name on the computer arc.ucalgary.ca.
Then the path follows after the ":" separator,
but it always starts at the user's home location, therefore the path
user.name@arc.ucalgary.ca:
simply means user's home directory. Clearly, the same directory can also be pointed as
user.name@arc.ucalgary.ca:/home/user.name/
but it is longer, contains redundant information (the name is given twice), and it means exactly the same thing.
Naturally, the file my_test.dat, could be referenced, when using the relative path as
user.name@arc.ucalgary.ca:my_data/my_test.dat
and using the absolute path as
user.name@arc.ucalgary.ca:/home/user.name/my_data/my_test.dat