1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
|
Returns or stores the elements contained in the [list][tdtl], [set][tdts] or
[sorted set][tdtss] at `key`.
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:
[tdtl]: /topics/data-types#lists
[tdts]: /topics/data-types#set
[tdtss]: /topics/data-types#sorted-sets
```
SORT mylist
```
Assuming `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 `!DESC` modifier:
```
SORT mylist DESC
```
When `mylist` contains string values and you want to sort them
lexicographically, use the `!ALPHA` modifier:
```
SORT mylist ALPHA
```
Redis is UTF-8 aware, assuming you correctly set the `!LC_COLLATE` environment
variable.
The number of returned elements can be limited using the `!LIMIT` modifier.
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
starting at `offset`.
The following example will return 10 elements of the sorted version of `mylist`,
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 `1`, `2` and `3` representing
unique IDs of objects stored in `object_1`, `object_2` and `object_3`.
When these objects have associated weights stored in `weight_1`, `weight_2` and
`weight_3`, `SORT` can be instructed to use these weights to sort `mylist` with
the following statement:
```
SORT mylist BY weight_*
```
The `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 (`1`, `2` and `3` in this example).
## Skip sorting the elements
The `!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 `!GET` option
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
(`object_1`, `object_2` and `object_3`).
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_*
```
The `!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 #
```
## Restrictions for using external keys
When enabling `Redis cluster-mode` there is no way to guarantee the existence of the external keys on the node which the command is processed on.
In this case, any use of `GET` or `BY` which reference external key pattern will cause the command to fail with an error.
Starting from Redis 7.0, any use of `GET` or `BY` which reference external key pattern will only be allowed in case the current user running the command has full key read permissions.
Full key read permissions can be set for the user by, for example, specifying `'%R~*'` or `'~*` with the relevant command access rules.
You can check the `ACL SETUSER` command manual for more information on setting ACL access rules.
If full key read permissions aren't set, the command will fail with an error.
## Storing the result of a SORT operation
By default, `SORT` returns the sorted elements to the client.
With the `!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
request.
When the key will timeout, an updated version of the cache can be created by
calling `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 `!BY` and `!GET`
It is possible to use `!BY` and `!GET` options against hash fields with the
following syntax:
```
SORT mylist BY weight_*->fieldname GET object_*->fieldname
```
The string `->` 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.
@return
@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.
|
