On: Autocomplete of entire Linux kernel using Redis's new ZRANGEBYLEX

[dave1010uk]

What's the performance of a MySQL LIKE query on a similar sized table (with an index)?

[gfv → dave1010uk]

It's not really possible to compare LIKE to ZRANGEBYLEX. LIKE in a general case is an infix search in all text in a column, while ZRANGEBYLEX is a special case of search of all values within a given range. ZRANGEBYLEX uses a lexicographic comparison, so one can only search for strings with a prefix between certain values.

[mappu → gfv]

MySQL can search by prefix by using only a trailing wildcard (LIKE "string%"), and then it will use an index if one is available.