The standard output of dysk is a table with a default set of columns and only the "normal looking" filesystems. You can modify it easily.

Columns

You can run dysk --list-cols for the list of all columns.

All columns

name default meaning
id mount point id
dev device id
filesystem filesystem
label label
type filesystem type
remote whether it's a remote filesystem
disk short tag of the underlying storage identified
used cumulated size of the occupied blocks
use graphical view of the use share
use_percent percentage of occupied blocks
free cumulated size of the available blocks
free_percent percentage of available blocks
size size of the volume
inodesfree available inodes
inodesused inodes used
inodes inodes use share, graphical
inodes_use_percent inodes use share, in percents
inodescount total number of inodes in the filesystem
mount mounting path

Choose columns

With the --cols launch argument, shortened as -c, you can change the displayed columns or their order.

The default set is defined for the casual usage of checking the available volumes and their filling level:

screen

With -c all, you may see all available columns, but that's normally too much for convenience:

screen

The most obvious use of the --cols argument is the explicit definition of the columns to display.

For example dysk -c label+use+size+disk+mount will show the label, use, size, disk, and mount columns, in that order:

screen

All the default columns (see table above) can be inserted with just default.

Here's adding the label at the start and the device id at the end, with dysk -c label+default+dev:

screen

If the --cols argument starts or ends with + or -, the default set of columns is implied. To add the device id and the share of inodes used to the default columns, you do dysk -c +dev+inodes:

screen

To preprend the label column before the default ones, use dysk -c label+:

screen

The - sign removes columns. And adding an already present column moves it to the end (there's never duplicates).

Here's removing the fs column and moving the type column to the end, with dysk -c -fs+type:

screen

Sort

With the --sort launch argument, shortened as -s, you can specify the order of displayed rows.

The argument's value must be either a column name, for example dysk -s dev, or a column name and a direction, for example dysk --sort size-desc.

The desc and asc directions can be abbreviated into d and a.

For example, sorting on the device id:

screen

Or sorting on the remaining free space, in descending order:

screen

CSV

With the --csv argument, you can ask dysk to output the table in CSV:

dysk --csv > mounts.csv

You may choose the separator with the --csv-separator argument.

Filters, sorting, and column selection work the same than for standard tables so you may do this:

dysk --csv -f 'size>100G' -c remote+default+inodes > mounts.csv

which would give something like this:

screen