Redis - 缓存数据的最佳方式

Question

I have 3 types of values.

Type1 - In millions
Type2 - In hundreds
Type3 - In tens

For each type1, all type2 should be processed by type3.

I can use set.

Type1#Type3 = {Type2 values}

Or

Type1#Type2 = {Type3 values}

If I keep Type3 as a key, more memory occupied by the set.

Or is there any better way to store this type data?

I evaluated hash which occupies more memory than set. I evaluated with ziplist compression for hash.

Any suggestions?

Answer 1

If you can tolerate splitting, here is another way:

First HashMap with key as a type1 value and value as a : separated list of type2 values linked to that type 1 value

type1_a => type2_b:type2_c..

type1_b => type2_k:type2_d..

Second HashMap with key as a type2 value and value as the corresponding type3 value

type2_a => type3a

type3_b => type3b

for type2_key in firstHashMap.get(type1_key).split(':') secondHasMap.get(type2_key)

Also, you can look into Lua script to avoid doing the above on client side.