There is also the
SORT_RO read-only variant of this command.
By default, sorting is numeric and elements are compared by their value interpreted as double precision floating point number. This is SORT in its simplest form:
mylist is a list of numbers, this command will return the same list
with the elements sorted from small to large.
In order to sort the numbers from large to small, use the
SORT mylist DESC
mylist contains string values and you want to sort them
lexicographically, use the
SORT mylist ALPHA
Redis is UTF-8 aware, assuming you correctly set the
The number of returned elements can be limited using the
This modifier takes the
offset argument, specifying the number of elements to
skip and the
count argument, specifying the number of elements to return from
The following example will return 10 elements of the sorted version of
starting at element 0 (
offset is zero-based):
SORT mylist LIMIT 0 10
Almost all modifiers can be used together. The following example will return the first 5 elements, lexicographically sorted in descending order:
SORT mylist LIMIT 0 5 ALPHA DESC
*Sorting by external keys
Sometimes you want to sort elements using external keys as weights to compare
instead of comparing the actual elements in the list, set or sorted set.
Let's say the list
mylist contains the elements
unique IDs of objects stored in
When these objects have associated weights stored in
weight_3, SORT can be instructed to use these weights to sort
the following statement:
SORT mylist BY weight_*
BY option takes a pattern (equal to
weight_* in this example) that is
used to generate the keys that are used for sorting.
These key names are obtained substituting the first occurrence of
* with the
actual value of the element in the list (
3 in this example).
*Skip sorting the elements
BY option can also take a non-existent key, which causes SORT to skip
the sorting operation.
This is useful if you want to retrieve external keys (see the
below) without the overhead of sorting.
SORT mylist BY nosort
*Retrieving external keys
Our previous example returns just the sorted IDs.
In some cases, it is more useful to get the actual objects instead of their IDs
Retrieving external keys based on the elements in a list, set or sorted set can
be done with the following command:
SORT mylist BY weight_* GET object_*
GET option can be used multiple times in order to get more keys for every
element of the original list, set or sorted set.
It is also possible to
GET the element itself using the special pattern
SORT mylist BY weight_* GET object_* GET #
*Storing the result of a SORT operation
By default, SORT returns the sorted elements to the client.
STORE option, the result will be stored as a list at the specified
key instead of being returned to the client.
SORT mylist BY weight_* STORE resultkey
An interesting pattern using
SORT ... STORE consists in associating an
EXPIRE timeout to the resulting key so that in applications where the result
of a SORT operation can be cached for some time.
Other clients will use the cached list instead of calling SORT for every
When the key will timeout, an updated version of the cache can be created by
SORT ... STORE again.
Note that for correctly implementing this pattern it is important to avoid multiple clients rebuilding the cache at the same time. Some kind of locking is needed here (for instance using SETNX).
*Using hashes in
It is possible to use
GET options against hash fields with the
SORT mylist BY weight_*->fieldname GET object_*->fieldname
-> is used to separate the key name from the hash field name.
The key is substituted as documented above, and the hash stored at the resulting
key is accessed to retrieve the specified hash field.
Array reply: without passing the
store option the command returns a list of sorted elements.
Integer reply: when the
store option is specified the command returns the number of sorted elements in the destination list.