Files
go-redis/sortedset_commands.go
Nedyalko Dyakov f813adcb85 feat(autopipeline): run N full-duplex engines on one client (#4022)
* feat(autopipeline): spike N full-duplex engines per client

SPIKE, not for merge. Answers whether the ~2x measured for
"multi-connection full duplex" is a property of N engines or an artifact
of the benchmark harness, which fakes multi-conn by building N separate
*redis.Client instances and round-robining across them. That holds six
sockets per client to do I/O on one -- 48 for eight wires -- and keeps
the routing outside the library, so key routing could not be tested.

NumShards>1 under FullDuplex now means N engines on ONE client: one pool,
one hook chain, one invalidation stream, N held wires. Commands route by
key hash, so same-key commands keep one wire and per-key order holds
without an Unordered opt-in. Keyless commands round-robin. A whole
FDPipelined batch pins to one engine, since splitting it would lose the
internal order a pipeline exists to guarantee.

The change is TWO-SITED, which is the finding. Relaxing the Validate
rejection alone silently turns full duplex OFF: fdOn is separately gated
on nShards == 1, so NumShards>1 fell through to half-duplex round-robin
shards, which are unordered, and a same-key write-then-read returned
redis.Nil. That is exactly what the deleted error warned about ("reject
it rather than silently falling back to half-duplex"). Both the
validation and the construction gate have to move together, with
nShards forced back to 1 so the shard array stays flusherless.

Measured at 1024 callers, 10 commands per Exec, one binary:

  proxy, 8 clients    1,686,080 ops/s   48 held sockets
  in-client, 8        1,791,369 ops/s   12 held sockets

Known gap before this could ship: engines lease from the pipeline pool
and DefaultPipelinePoolSize is 10, so more than ~10 engines spill to the
main pool. A real feature must size that pool from the engine count.
cmdFirstKey is also deliberately crude (args[1]) rather than using
declared key positions.

* refactor(autopipeline): document FD sharding, guard the pool

Drops the exploratory framing from the multi-connection full-duplex code
and closes two of its three known gaps.

Key extraction no longer assumes args[1]. It calls
cmdFirstKeyPosWithInfo(cmd, nil), the same helper client-side caching
uses, so keyless commands, the static command table, and eval/evalsha
variants whose key position depends on the runtime numkeys argument all
route correctly instead of hashing a non-key argument.

Engine count is now checked against the pipeline pool. Each engine holds
one pooled connection for as long as it runs, and DefaultPipelinePoolSize
is 10, so 11 or more engines left the surplus permanently spilling onto
the main pool and competing with ordinary commands. Construction fails
with a message naming PipelinePoolSize instead.

The two-sited rule is now stated at both sites: relaxing the Validate
rejection without relaxing the fdOn gate leaves full duplex silently OFF
for NumShards>1, so commands fall through to unordered half-duplex shards
and a same-key write-then-read can read a stale value. Tests cover it
from both directions -- reply correlation and per-key ordering across
eight engines, and Config() reporting FullDuplex true rather than having
fallen back.

Still open, and why this stays draft: NumShards:0 means one engine and
sizing is not automatic. A fixed count above 1 costs low-concurrency
callers about 30% (eight engines at eight callers), so the count should
be derived from observed load, roughly one wire per 256-512 callers.

* fix(autopipeline): fix FD engine routing edges

- The NumShards>1 ordering check exempted any requested FullDuplex.
  Without a pipeline pool full duplex does not engage and construction
  falls back to round-robin shards, which need Unordered. The exemption
  now follows the effective state.
- Keys are read with stringArg, as Ring does, so []byte, *string and
  numeric keys route like the string form, and an empty key routes by
  hash instead of round-robin.
- Engine index uses uint32 modulo; converting to int first goes
  negative on 32-bit platforms.
- Document that N engines keep order per first key only, so multi-key
  commands are not ordered across engines, and that NumShards>1 under
  full duplex is new, so the mode is opt-in.

* docs(autopipeline): FD engine ordering and window

Document that with NumShards > 1 keyless commands, including ones that
change every key (FLUSHDB, SWAPDB), round-robin and are not ordered
against keyed commands; that keys route through stringArg like Ring and
the cluster slot hash; and that FullDuplexWindow applies per engine.

* docs(autopipeline): pool check is per autopipeliner

The NumShards pipeline-pool check counts only the autopipeliner being
built. The blocking and async faces and WithTimeout clones share one
pipeline pool, so it must hold the sum of their engines. Say so; a
client-wide engine count is a follow-up.

* fix(command): route raw FCALL by its first key

FCALL has the EVAL layout (name, function, numkeys, keys...), but the key
resolver only special-cased EVAL. The typed FCall sets its key position;
a raw Do/Process FCALL fell back to position 1 and hashed the function
name. That misrouted it on Ring (wrong shard), on cluster (a MOVED away)
and across full-duplex engines (a following GET of its key could run
first). A zero-key FCALL now resolves as keyless, like a zero-key EVAL.

* docs(autopipeline): document multi-engine NumShards

The exported FullDuplex and NumShards docs still described a single
full-duplex stream and said NumShards > 1 needs Unordered. Describe the
multi-engine mode: engines routed by first key, per-first-key order
only, per-engine window, and the pipeline pool sizing.

* fix(autopipeline): keep FD batches on their keys' engine

With several full-duplex engines, FDPipelined sent the whole batch to
its first command's engine. A later command in the batch whose key
hashes elsewhere could then run ahead of an earlier, unawaited command
for that key on that key's engine.

A batch now rides an engine only when every keyed command hashes to it
(keyless commands ride along). Otherwise FDPipelined returns
ErrFDPipelineSpansEngines and Pipeline() runs an ordinary pipeline.

* docs(autopipeline): note key types in engine routing

With several full-duplex engines a command is routed by a hash of its
first key in string form (fmt for non-string arguments, as Ring does),
so the same Redis key passed as different Go types (true and "1") can
land on different engines. Say so on NumShards.

* test(autopipeline): use the configured test address

The full-duplex shard tests dialed 127.0.0.1:6379 directly and skipped
under the REDIS_PORT override. Use apTestAddr and internalTestRedisAddr
like the other autopipeline tests.

* fix(command): route raw XGROUP and ZUNION by key

The raw-key resolver fell back to args[1] for XGROUP, whose key follows
the subcommand, and for ZUNION, ZINTER and ZDIFF, whose key follows
numkeys. With several full-duplex engines a raw command could then run
on a different engine from earlier writes to the same key, and Ring
picked the wrong shard. The typed constructors already set position 2.

* fix(command): route raw keyed commands by key

The raw-key resolver fell back to args[1] for SDIFFCARD, SUNIONCARD,
TS.NRANGE and TS.NREVRANGE (key after numkeys) and HIMPORT SET (key
after the subcommand), where the typed constructors set position 2.

HSCAN, SSCAN and ZSCAN with a hash tag in MATCH routed by the pattern,
a hint meant for the keyless SCAN. The key decides the node: a cluster
scan took a MOVED hop, a Ring scan read the wrong shard, and a
multi-engine full-duplex scan could pass earlier writes to its key.

* docs(autopipeline): note the cost of FD engines

NumShards is a fixed engine count, and several engines fragment small
batches. Give the measured trade on NumShards: 8 engines were 30%
slower than 1 at 8 callers, broke even at about 128, and were 2.2x
faster at 1024.
2026-10-02 16:49:25 +03:00

821 lines
24 KiB
Go

package redis
import (
"context"
"errors"
"strings"
"time"
)
type SortedSetCmdable interface {
BZPopMax(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd
BZPopMin(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd
BZMPop(ctx context.Context, timeout time.Duration, order string, count int64, keys ...string) *ZSliceWithKeyCmd
ZAdd(ctx context.Context, key string, members ...Z) *IntCmd
ZAddLT(ctx context.Context, key string, members ...Z) *IntCmd
ZAddGT(ctx context.Context, key string, members ...Z) *IntCmd
ZAddNX(ctx context.Context, key string, members ...Z) *IntCmd
ZAddXX(ctx context.Context, key string, members ...Z) *IntCmd
ZAddArgs(ctx context.Context, key string, args ZAddArgs) *IntCmd
ZAddArgsIncr(ctx context.Context, key string, args ZAddArgs) *FloatCmd
ZCard(ctx context.Context, key string) *IntCmd
ZCount(ctx context.Context, key, min, max string) *IntCmd
ZLexCount(ctx context.Context, key, min, max string) *IntCmd
ZIncrBy(ctx context.Context, key string, increment float64, member string) *FloatCmd
ZInter(ctx context.Context, store *ZStore) *StringSliceCmd
ZInterWithScores(ctx context.Context, store *ZStore) *ZSliceCmd
ZInterCard(ctx context.Context, limit int64, keys ...string) *IntCmd
ZInterStore(ctx context.Context, destination string, store *ZStore) *IntCmd
ZMPop(ctx context.Context, order string, count int64, keys ...string) *ZSliceWithKeyCmd
ZMScore(ctx context.Context, key string, members ...string) *FloatSliceCmd
ZPopMax(ctx context.Context, key string, count ...int64) *ZSliceCmd
ZPopMin(ctx context.Context, key string, count ...int64) *ZSliceCmd
ZRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd
ZRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd
ZRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
ZRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
ZRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd
ZRangeArgs(ctx context.Context, z ZRangeArgs) *StringSliceCmd
ZRangeArgsWithScores(ctx context.Context, z ZRangeArgs) *ZSliceCmd
ZRangeStore(ctx context.Context, dst string, z ZRangeArgs) *IntCmd
ZRank(ctx context.Context, key, member string) *IntCmd
ZRankWithScore(ctx context.Context, key, member string) *RankWithScoreCmd
ZRem(ctx context.Context, key string, members ...interface{}) *IntCmd
ZRemRangeByRank(ctx context.Context, key string, start, stop int64) *IntCmd
ZRemRangeByScore(ctx context.Context, key, min, max string) *IntCmd
ZRemRangeByLex(ctx context.Context, key, min, max string) *IntCmd
ZRevRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd
ZRevRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd
ZRevRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
ZRevRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd
ZRevRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd
ZRevRank(ctx context.Context, key, member string) *IntCmd
ZRevRankWithScore(ctx context.Context, key, member string) *RankWithScoreCmd
ZScore(ctx context.Context, key, member string) *FloatCmd
ZUnionStore(ctx context.Context, dest string, store *ZStore) *IntCmd
ZRandMember(ctx context.Context, key string, count int) *StringSliceCmd
ZRandMemberWithScores(ctx context.Context, key string, count int) *ZSliceCmd
ZUnion(ctx context.Context, store ZStore) *StringSliceCmd
ZUnionWithScores(ctx context.Context, store ZStore) *ZSliceCmd
ZDiff(ctx context.Context, keys ...string) *StringSliceCmd
ZDiffWithScores(ctx context.Context, keys ...string) *ZSliceCmd
ZDiffStore(ctx context.Context, destination string, keys ...string) *IntCmd
ZScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd
}
// BZPopMax Redis `BZPOPMAX key [key ...] timeout` command.
func (c cmdable) BZPopMax(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd {
args := make([]interface{}, 1+len(keys)+1)
args[0] = "bzpopmax"
for i, key := range keys {
args[1+i] = key
}
args[len(args)-1] = formatSec(ctx, timeout)
cmd := NewZWithKeyCmd(ctx, args...)
cmd.setReadTimeout(timeout)
_ = c(ctx, cmd)
return cmd
}
// BZPopMin Redis `BZPOPMIN key [key ...] timeout` command.
func (c cmdable) BZPopMin(ctx context.Context, timeout time.Duration, keys ...string) *ZWithKeyCmd {
args := make([]interface{}, 1+len(keys)+1)
args[0] = "bzpopmin"
for i, key := range keys {
args[1+i] = key
}
args[len(args)-1] = formatSec(ctx, timeout)
cmd := NewZWithKeyCmd(ctx, args...)
cmd.setReadTimeout(timeout)
_ = c(ctx, cmd)
return cmd
}
// BZMPop is the blocking variant of ZMPOP.
// When any of the sorted sets contains elements, this command behaves exactly like ZMPOP.
// When all sorted sets are empty, Redis will block the connection until another client adds members to one of the keys or until the timeout elapses.
// A timeout of zero can be used to block indefinitely.
// example: client.BZMPop(ctx, 0,"max", 1, "set")
func (c cmdable) BZMPop(ctx context.Context, timeout time.Duration, order string, count int64, keys ...string) *ZSliceWithKeyCmd {
args := make([]interface{}, 3+len(keys), 6+len(keys))
args[0] = "bzmpop"
args[1] = formatSec(ctx, timeout)
args[2] = len(keys)
for i, key := range keys {
args[3+i] = key
}
args = append(args, strings.ToLower(order), "count", count)
cmd := NewZSliceWithKeyCmd(ctx, args...)
// The key follows numkeys, not args[1]: route by it.
if len(keys) > 0 {
cmd.SetFirstKeyPos(3)
}
cmd.setReadTimeout(timeout)
_ = c(ctx, cmd)
return cmd
}
// ZAddArgs WARN: The GT, LT and NX options are mutually exclusive.
type ZAddArgs struct {
NX bool
XX bool
LT bool
GT bool
Ch bool
Members []Z
}
func (c cmdable) zAddArgs(key string, args ZAddArgs, incr bool) []interface{} {
a := make([]interface{}, 0, 6+2*len(args.Members))
a = append(a, "zadd", key)
// The GT, LT and NX options are mutually exclusive.
if args.NX {
a = append(a, "nx")
} else {
if args.XX {
a = append(a, "xx")
}
if args.GT {
a = append(a, "gt")
} else if args.LT {
a = append(a, "lt")
}
}
if args.Ch {
a = append(a, "ch")
}
if incr {
a = append(a, "incr")
}
for _, m := range args.Members {
a = append(a, m.Score)
a = append(a, m.Member)
}
return a
}
func (c cmdable) ZAddArgs(ctx context.Context, key string, args ZAddArgs) *IntCmd {
cmd := NewIntCmd(ctx, c.zAddArgs(key, args, false)...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZAddArgsIncr(ctx context.Context, key string, args ZAddArgs) *FloatCmd {
cmd := NewFloatCmd(ctx, c.zAddArgs(key, args, true)...)
_ = c(ctx, cmd)
return cmd
}
// ZAdd Redis `ZADD key score member [score member ...]` command.
func (c cmdable) ZAdd(ctx context.Context, key string, members ...Z) *IntCmd {
return c.ZAddArgs(ctx, key, ZAddArgs{
Members: members,
})
}
// ZAddLT Redis `ZADD key LT score member [score member ...]` command.
func (c cmdable) ZAddLT(ctx context.Context, key string, members ...Z) *IntCmd {
return c.ZAddArgs(ctx, key, ZAddArgs{
LT: true,
Members: members,
})
}
// ZAddGT Redis `ZADD key GT score member [score member ...]` command.
func (c cmdable) ZAddGT(ctx context.Context, key string, members ...Z) *IntCmd {
return c.ZAddArgs(ctx, key, ZAddArgs{
GT: true,
Members: members,
})
}
// ZAddNX Redis `ZADD key NX score member [score member ...]` command.
func (c cmdable) ZAddNX(ctx context.Context, key string, members ...Z) *IntCmd {
return c.ZAddArgs(ctx, key, ZAddArgs{
NX: true,
Members: members,
})
}
// ZAddXX Redis `ZADD key XX score member [score member ...]` command.
func (c cmdable) ZAddXX(ctx context.Context, key string, members ...Z) *IntCmd {
return c.ZAddArgs(ctx, key, ZAddArgs{
XX: true,
Members: members,
})
}
func (c cmdable) ZCard(ctx context.Context, key string) *IntCmd {
cmd := NewIntCmd(ctx, "zcard", key)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZCount(ctx context.Context, key, min, max string) *IntCmd {
cmd := NewIntCmd(ctx, "zcount", key, min, max)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZLexCount(ctx context.Context, key, min, max string) *IntCmd {
cmd := NewIntCmd(ctx, "zlexcount", key, min, max)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZIncrBy(ctx context.Context, key string, increment float64, member string) *FloatCmd {
cmd := NewFloatCmd(ctx, "zincrby", key, increment, member)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZInterStore(ctx context.Context, destination string, store *ZStore) *IntCmd {
args := make([]interface{}, 0, 3+store.len())
args = append(args, "zinterstore", destination, len(store.Keys))
args = store.appendArgs(args)
cmd := NewIntCmd(ctx, args...)
cmd.SetFirstKeyPos(3)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZInter(ctx context.Context, store *ZStore) *StringSliceCmd {
args := make([]interface{}, 0, 2+store.len())
args = append(args, "zinter", len(store.Keys))
args = store.appendArgs(args)
cmd := NewStringSliceCmd(ctx, args...)
cmd.SetFirstKeyPos(2)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZInterWithScores(ctx context.Context, store *ZStore) *ZSliceCmd {
args := make([]interface{}, 0, 3+store.len())
args = append(args, "zinter", len(store.Keys))
args = store.appendArgs(args)
args = append(args, "withscores")
cmd := NewZSliceCmd(ctx, args...)
cmd.SetFirstKeyPos(2)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZInterCard(ctx context.Context, limit int64, keys ...string) *IntCmd {
numKeys := len(keys)
args := make([]interface{}, 4+numKeys)
args[0] = "zintercard"
args[1] = numKeys
for i, key := range keys {
args[2+i] = key
}
args[2+numKeys] = "limit"
args[3+numKeys] = limit
cmd := NewIntCmd(ctx, args...)
// The key follows numkeys, not args[1]: route by it.
if len(keys) > 0 {
cmd.SetFirstKeyPos(2)
}
_ = c(ctx, cmd)
return cmd
}
// ZMPop Pops one or more elements with the highest or lowest score from the first non-empty sorted set key from the list of provided key names.
// direction: "max" (highest score) or "min" (lowest score), count: > 0
// example: client.ZMPop(ctx, "max", 5, "set1", "set2")
func (c cmdable) ZMPop(ctx context.Context, order string, count int64, keys ...string) *ZSliceWithKeyCmd {
args := make([]interface{}, 2+len(keys), 5+len(keys))
args[0] = "zmpop"
args[1] = len(keys)
for i, key := range keys {
args[2+i] = key
}
args = append(args, strings.ToLower(order), "count", count)
cmd := NewZSliceWithKeyCmd(ctx, args...)
// The key follows numkeys, not args[1]: route by it.
if len(keys) > 0 {
cmd.SetFirstKeyPos(2)
}
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZMScore(ctx context.Context, key string, members ...string) *FloatSliceCmd {
args := make([]interface{}, 2+len(members))
args[0] = "zmscore"
args[1] = key
for i, member := range members {
args[2+i] = member
}
cmd := NewFloatSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZPopMax(ctx context.Context, key string, count ...int64) *ZSliceCmd {
args := []interface{}{
"zpopmax",
key,
}
switch len(count) {
case 0:
break
case 1:
args = append(args, count[0])
default:
cmd := NewZSliceCmd(ctx)
cmd.SetErr(errors.New("too many arguments"))
return cmd
}
cmd := NewZSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZPopMin(ctx context.Context, key string, count ...int64) *ZSliceCmd {
args := []interface{}{
"zpopmin",
key,
}
switch len(count) {
case 0:
break
case 1:
args = append(args, count[0])
default:
cmd := NewZSliceCmd(ctx)
cmd.SetErr(errors.New("too many arguments"))
return cmd
}
cmd := NewZSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
// ZRangeArgs is all the options of the ZRange command.
// In version> 6.2.0, you can replace the(cmd):
//
// ZREVRANGE,
// ZRANGEBYSCORE,
// ZREVRANGEBYSCORE,
// ZRANGEBYLEX,
// ZREVRANGEBYLEX.
//
// Please pay attention to your redis-server version.
//
// Rev, ByScore, ByLex and Offset+Count options require redis-server 6.2.0 and higher.
type ZRangeArgs struct {
Key string
// When the ByScore option is provided, the open interval(exclusive) can be set.
// By default, the score intervals specified by <Start> and <Stop> are closed (inclusive).
// It is similar to the deprecated(6.2.0+) ZRangeByScore command.
// For example:
// ZRangeArgs{
// Key: "example-key",
// Start: "(3",
// Stop: 8,
// ByScore: true,
// }
// cmd: "ZRange example-key (3 8 ByScore" (3 < score <= 8).
//
// When the Rev option is also provided, <Start> should be the higher score value and
// <Stop> should be the lower score value (i.e. reversed order):
// ZRangeArgs{
// Key: "example-key",
// Start: 8,
// Stop: "(3",
// ByScore: true,
// Rev: true,
// }
// cmd: "ZRange example-key 8 (3 ByScore Rev" (8 >= score > 3, in reverse order).
//
// For the ByLex option, it is similar to the deprecated(6.2.0+) ZRangeByLex command.
// You can set the <Start> and <Stop> options as follows:
// ZRangeArgs{
// Key: "example-key",
// Start: "[abc",
// Stop: "(def",
// ByLex: true,
// }
// cmd: "ZRange example-key [abc (def ByLex"
//
// When the Rev option is also provided, <Start> should be the lexicographically higher
// value and <Stop> should be the lower value:
// ZRangeArgs{
// Key: "example-key",
// Start: "(def",
// Stop: "[abc",
// ByLex: true,
// Rev: true,
// }
// cmd: "ZRange example-key (def [abc ByLex Rev"
//
// For normal cases (ByScore==false && ByLex==false), <Start> and <Stop> should be set to the index range (int).
// You can read the documentation for more information: https://redis.io/commands/zrange
Start interface{}
Stop interface{}
// The ByScore and ByLex options are mutually exclusive.
ByScore bool
ByLex bool
Rev bool
// limit offset count.
Offset int64
Count int64
}
func (z ZRangeArgs) appendArgs(args []interface{}) []interface{} {
args = append(args, z.Key, z.Start, z.Stop)
if z.ByScore {
args = append(args, "byscore")
} else if z.ByLex {
args = append(args, "bylex")
}
if z.Rev {
args = append(args, "rev")
}
if z.Offset != 0 || z.Count != 0 {
args = append(args, "limit", z.Offset, z.Count)
}
return args
}
func (c cmdable) ZRangeArgs(ctx context.Context, z ZRangeArgs) *StringSliceCmd {
args := make([]interface{}, 0, 9)
args = append(args, "zrange")
args = z.appendArgs(args)
cmd := NewStringSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRangeArgsWithScores(ctx context.Context, z ZRangeArgs) *ZSliceCmd {
args := make([]interface{}, 0, 10)
args = append(args, "zrange")
args = z.appendArgs(args)
args = append(args, "withscores")
cmd := NewZSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd {
return c.ZRangeArgs(ctx, ZRangeArgs{
Key: key,
Start: start,
Stop: stop,
})
}
func (c cmdable) ZRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd {
return c.ZRangeArgsWithScores(ctx, ZRangeArgs{
Key: key,
Start: start,
Stop: stop,
})
}
type ZRangeBy struct {
Min, Max string
Offset, Count int64
}
func (c cmdable) zRangeBy(ctx context.Context, zcmd, key string, opt *ZRangeBy, withScores bool) *StringSliceCmd {
args := []interface{}{zcmd, key, opt.Min, opt.Max}
if withScores {
args = append(args, "withscores")
}
if opt.Offset != 0 || opt.Count != 0 {
args = append(
args,
"limit",
opt.Offset,
opt.Count,
)
}
cmd := NewStringSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
// ZRangeByScore returns members in a sorted set within a range of scores.
//
// Deprecated: Use ZRangeArgs with ByScore option instead as of Redis 6.2.0.
func (c cmdable) ZRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
return c.zRangeBy(ctx, "zrangebyscore", key, opt, false)
}
// ZRangeByLex returns members in a sorted set within a lexicographical range.
//
// Deprecated: Use ZRangeArgs with ByLex option instead as of Redis 6.2.0.
func (c cmdable) ZRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
return c.zRangeBy(ctx, "zrangebylex", key, opt, false)
}
func (c cmdable) ZRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd {
args := []interface{}{"zrangebyscore", key, opt.Min, opt.Max, "withscores"}
if opt.Offset != 0 || opt.Count != 0 {
args = append(
args,
"limit",
opt.Offset,
opt.Count,
)
}
cmd := NewZSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRangeStore(ctx context.Context, dst string, z ZRangeArgs) *IntCmd {
args := make([]interface{}, 0, 10)
args = append(args, "zrangestore", dst)
args = z.appendArgs(args)
cmd := NewIntCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRank(ctx context.Context, key, member string) *IntCmd {
cmd := NewIntCmd(ctx, "zrank", key, member)
_ = c(ctx, cmd)
return cmd
}
// ZRankWithScore according to the Redis documentation, if member does not exist
// in the sorted set or key does not exist, it will return a redis.Nil error.
func (c cmdable) ZRankWithScore(ctx context.Context, key, member string) *RankWithScoreCmd {
cmd := NewRankWithScoreCmd(ctx, "zrank", key, member, "withscore")
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRem(ctx context.Context, key string, members ...interface{}) *IntCmd {
args := make([]interface{}, 2, 2+len(members))
args[0] = "zrem"
args[1] = key
args = appendArgs(args, members)
cmd := NewIntCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRemRangeByRank(ctx context.Context, key string, start, stop int64) *IntCmd {
cmd := NewIntCmd(
ctx,
"zremrangebyrank",
key,
start,
stop,
)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRemRangeByScore(ctx context.Context, key, min, max string) *IntCmd {
cmd := NewIntCmd(ctx, "zremrangebyscore", key, min, max)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRemRangeByLex(ctx context.Context, key, min, max string) *IntCmd {
cmd := NewIntCmd(ctx, "zremrangebylex", key, min, max)
_ = c(ctx, cmd)
return cmd
}
// ZRevRange returns members in a sorted set within a range of indexes in reverse order.
//
// Deprecated: Use ZRangeArgs with Rev option instead as of Redis 6.2.0.
func (c cmdable) ZRevRange(ctx context.Context, key string, start, stop int64) *StringSliceCmd {
cmd := NewStringSliceCmd(ctx, "zrevrange", key, start, stop)
_ = c(ctx, cmd)
return cmd
}
// ZRevRangeWithScores according to the Redis documentation, if member does not exist
// in the sorted set or key does not exist, it will return a redis.Nil error.
func (c cmdable) ZRevRangeWithScores(ctx context.Context, key string, start, stop int64) *ZSliceCmd {
cmd := NewZSliceCmd(ctx, "zrevrange", key, start, stop, "withscores")
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) zRevRangeBy(ctx context.Context, zcmd, key string, opt *ZRangeBy) *StringSliceCmd {
args := []interface{}{zcmd, key, opt.Max, opt.Min}
if opt.Offset != 0 || opt.Count != 0 {
args = append(
args,
"limit",
opt.Offset,
opt.Count,
)
}
cmd := NewStringSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
// ZRevRangeByScore returns members in a sorted set within a range of scores in reverse order.
//
// Deprecated: Use ZRangeArgs with Rev and ByScore options instead as of Redis 6.2.0.
func (c cmdable) ZRevRangeByScore(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
return c.zRevRangeBy(ctx, "zrevrangebyscore", key, opt)
}
// ZRevRangeByLex returns members in a sorted set within a lexicographical range in reverse order.
//
// Deprecated: Use ZRangeArgs with Rev and ByLex options instead as of Redis 6.2.0.
func (c cmdable) ZRevRangeByLex(ctx context.Context, key string, opt *ZRangeBy) *StringSliceCmd {
return c.zRevRangeBy(ctx, "zrevrangebylex", key, opt)
}
func (c cmdable) ZRevRangeByScoreWithScores(ctx context.Context, key string, opt *ZRangeBy) *ZSliceCmd {
args := []interface{}{"zrevrangebyscore", key, opt.Max, opt.Min, "withscores"}
if opt.Offset != 0 || opt.Count != 0 {
args = append(
args,
"limit",
opt.Offset,
opt.Count,
)
}
cmd := NewZSliceCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRevRank(ctx context.Context, key, member string) *IntCmd {
cmd := NewIntCmd(ctx, "zrevrank", key, member)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZRevRankWithScore(ctx context.Context, key, member string) *RankWithScoreCmd {
cmd := NewRankWithScoreCmd(ctx, "zrevrank", key, member, "withscore")
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZScore(ctx context.Context, key, member string) *FloatCmd {
cmd := NewFloatCmd(ctx, "zscore", key, member)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZUnion(ctx context.Context, store ZStore) *StringSliceCmd {
args := make([]interface{}, 0, 2+store.len())
args = append(args, "zunion", len(store.Keys))
args = store.appendArgs(args)
cmd := NewStringSliceCmd(ctx, args...)
cmd.SetFirstKeyPos(2)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZUnionWithScores(ctx context.Context, store ZStore) *ZSliceCmd {
args := make([]interface{}, 0, 3+store.len())
args = append(args, "zunion", len(store.Keys))
args = store.appendArgs(args)
args = append(args, "withscores")
cmd := NewZSliceCmd(ctx, args...)
cmd.SetFirstKeyPos(2)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZUnionStore(ctx context.Context, dest string, store *ZStore) *IntCmd {
args := make([]interface{}, 0, 3+store.len())
args = append(args, "zunionstore", dest, len(store.Keys))
args = store.appendArgs(args)
cmd := NewIntCmd(ctx, args...)
cmd.SetFirstKeyPos(3)
_ = c(ctx, cmd)
return cmd
}
// ZRandMember redis-server version >= 6.2.0.
func (c cmdable) ZRandMember(ctx context.Context, key string, count int) *StringSliceCmd {
cmd := NewStringSliceCmd(ctx, "zrandmember", key, count)
_ = c(ctx, cmd)
return cmd
}
// ZRandMemberWithScores redis-server version >= 6.2.0.
func (c cmdable) ZRandMemberWithScores(ctx context.Context, key string, count int) *ZSliceCmd {
cmd := NewZSliceCmd(ctx, "zrandmember", key, count, "withscores")
_ = c(ctx, cmd)
return cmd
}
// ZDiff redis-server version >= 6.2.0.
func (c cmdable) ZDiff(ctx context.Context, keys ...string) *StringSliceCmd {
args := make([]interface{}, 2+len(keys))
args[0] = "zdiff"
args[1] = len(keys)
for i, key := range keys {
args[i+2] = key
}
cmd := NewStringSliceCmd(ctx, args...)
cmd.SetFirstKeyPos(2)
_ = c(ctx, cmd)
return cmd
}
// ZDiffWithScores redis-server version >= 6.2.0.
func (c cmdable) ZDiffWithScores(ctx context.Context, keys ...string) *ZSliceCmd {
args := make([]interface{}, 3+len(keys))
args[0] = "zdiff"
args[1] = len(keys)
for i, key := range keys {
args[i+2] = key
}
args[len(keys)+2] = "withscores"
cmd := NewZSliceCmd(ctx, args...)
cmd.SetFirstKeyPos(2)
_ = c(ctx, cmd)
return cmd
}
// ZDiffStore redis-server version >=6.2.0.
func (c cmdable) ZDiffStore(ctx context.Context, destination string, keys ...string) *IntCmd {
args := make([]interface{}, 0, 3+len(keys))
args = append(args, "zdiffstore", destination, len(keys))
for _, key := range keys {
args = append(args, key)
}
cmd := NewIntCmd(ctx, args...)
_ = c(ctx, cmd)
return cmd
}
func (c cmdable) ZScan(ctx context.Context, key string, cursor uint64, match string, count int64) *ScanCmd {
args := []interface{}{"zscan", key, cursor}
if match != "" {
args = append(args, "match", match)
}
if count > 0 {
args = append(args, "count", count)
}
cmd := NewScanCmd(ctx, c, args...)
_ = c(ctx, cmd)
return cmd
}
// Z represents sorted set member.
type Z struct {
Score float64
Member interface{}
}
// ZWithKey represents sorted set member including the name of the key where it was popped.
type ZWithKey struct {
Z
Key string
}
// ZStore is used as an arg to ZInter/ZInterStore and ZUnion/ZUnionStore.
type ZStore struct {
Keys []string
Weights []float64
// Can be SUM, MIN, MAX or COUNT.
Aggregate string
}
func (z ZStore) len() (n int) {
n = len(z.Keys)
if len(z.Weights) > 0 {
n += 1 + len(z.Weights)
}
if z.Aggregate != "" {
n += 2
}
return n
}
func (z ZStore) appendArgs(args []interface{}) []interface{} {
for _, key := range z.Keys {
args = append(args, key)
}
if len(z.Weights) > 0 {
args = append(args, "weights")
for _, weights := range z.Weights {
args = append(args, weights)
}
}
if z.Aggregate != "" {
args = append(args, "aggregate", z.Aggregate)
}
return args
}