一次数据操作耗时 = 网络 RTT + Redis 执行时间 + 客户端处理时间
                  ~1ms           ~0.01ms      ~0.01ms

# 适用：无依赖的批量操作
def batch_set_user(redis, users):
    """批量写入用户数据"""
    pipe = redis.pipeline()
    for uid, data in users.items():
        pipe.set(f"user:{uid}:name", data['name'])
        pipe.set(f"user:{uid}:age", data['age'])
        pipe.set(f"user:{uid}:city", data['city'])
    pipe.execute()  # 3*N 次 SET = 1 次网络往返

# 适用：多条 GET/SET 操作
# 不用 Pipeline，直接使用原生命令
r.mset({
    'user:1:name': 'Alice',
    'user:1:age': '30',
    'user:1:city': 'Beijing'
})

# 批量读取
names = r.mget('user:1:name', 'user:2:name', 'user:3:name')

# 适用：有逻辑依赖的批量操作
script = """
for i, key in ipairs(KEYS) do
    redis.call('SET', key, ARGV[i])
end
return #KEYS
"""
r.eval(script, 3, 'k1', 'k2', 'k3', 'v1', 'v2', 'v3')

# 适用：Redis 在本地部署
r = redis.Redis(unix_socket_path='/var/run/redis/redis.sock')

import redis

r = redis.Redis(host='localhost', port=6379)

# ❌ 错误做法：逐条导入
import time
start = time.time()
for i in range(1000000):
    r.set(f"data:{i}", f"value:{i}")
print(f"逐条导入: {time.time() - start:.2f}s")
# 约 280 秒 (内网环境)

# ✅ 正确做法：Pipeline 分批导入
start = time.time()
pipe = r.pipeline()
BATCH_SIZE = 1000
for i in range(1000000):
    pipe.set(f"data:{i}", f"value:{i}")
    if (i + 1) % BATCH_SIZE == 0:
        pipe.execute()
        pipe = r.pipeline()
# 最后一轮
pipe.execute()
print(f"Pipeline 导入: {time.time() - start:.2f}s")
# 约 8 秒 (内网环境)

# ❌ 分开操作
for uid in [1, 2, 3]:
    r.hset(f"user:{uid}", "online", "1")

# ✅ 合并操作
pipe = r.pipeline()
for uid in [1, 2, 3]:
    pipe.hset(f"user:{uid}", "online", "1")
pipe.execute()

# ❌ 多个命令
r.incr('views:article:1')
r.expire('views:article:1', 3600)

# ✅ Lua 脚本合并为一个发送
script = """
redis.call('INCR', KEYS[1])
redis.call('EXPIRE', KEYS[1], ARGV[1])
"""
r.eval(script, 1, 'views:article:1', 3600)

# ❸ 使用连接池复用连接
pool = redis.ConnectionPool(host='localhost', port=6379, max_connections=50)
r = redis.Redis(connection_pool=pool)
# 每次操作都复用已有连接，减少 TCP 握手