我应该在MySQL二进制数据类型列上使用哪个索引

Question

I am writing a simple tool to check duplicate files(ie files having same data). The mechanism is to generate hashes for each file using sha-512 algorithm and then store these hashes in MYSQL database. I store hashes in binary(64) unique not null column. Each row will have a unique binary hash and used to check file is duplicate or not.

-- My questions are --

1. Can I use indexes on binary column, my default table collation is latin1 - default collation?

2. Which Indexing mechanism should I use Btree or Hash, for getting high performance? I need to update or add 100 of rows per seconds.

3. What other things should I take care of to get best performance?

Answer 1

1. Can I use indexes on binary column, my default table collation is latin1 - default collation?

   Yes, you can; collation is only relevant for character datatypes, not binary datatypes (it defines how characters should be ordered)—also, be aware that latin1 is a character encoding , not a collation.

2. Which Indexing mechanism should I use Btree or Hash, for getting high performance? I need to update or add 100 of rows per seconds.

   Note that hash indexes are only available with the MEMORY and NDB storage engines, so you may not even have a choice.

   In any event, either would typically be able to meet your performance criteria—although for this particular application I see no benefit from using B-Tree (which is ordered), whereas Hash would give better performance. Therefore, if you have the choice, you may as well use Hash.

   See Comparison of B-Tree and Hash Indexes for more information.

3. What other things should I take care of to get best performance?

   Depends on your definition of "best performance" and your environment. In general, remember Knuth's maxim " premature optimisation is the root of all evil ": that is, only optimise when you know that there will be a problem with the simplest approach.