sds / mock_redis / 16382133615

98%

push

github

Faster hmget when fetching a lot of fields (#330)

The previous version of `hmget` calls `hget` for each field being
fetched, but that results in a lot of redundant calls to `with_hash_at`
(and in turn, `clean_up_empties_at`, which is not especially fast).

Change it to just use a single `with_hash_at` call.


```ruby
days = 1.upto(365).map(&:to_s)
redis = MockRedis.new
Benchmark.ips do |x|
  x.report("v1") { redis.hmget("views_by_year:2025", *days) }
  x.report("v2") { redis.hmget2("views_by_year:2025", *days) }
  x.compare!
end

# ruby 3.4.3 (2025-04-14 revision d0b7e5b6a0) +PRISM [arm64-darwin24]
# Warming up --------------------------------------
#                 v1    60.000 i/100ms
#                 v2   796.000 i/100ms
# Calculating -------------------------------------
#                 v1    608.372 (± 1.3%) i/s    (1.64 ms/i) -      3.060k in   5.030757s
#                 v2      8.212k (± 8.7%) i/s  (121.78 μs/i) -     41.392k in   5.087108s
# 
# Comparison:
#                 v2:     8211.7 i/s
#                 v1:      608.4 i/s - 13.50x  slower
```

1 of 1 new or added line in 1 file covered. (100.0%)

1864 of 1897 relevant lines covered (98.26%)

1624.76 hits per line