I've got a huge table with something like 8 300 000 rows (won't be edit nor delete ever).
My first column look something similar P300-4312B_X16_S
and the entry isn't unique so I use a regular INDEX on this field.
However, MySQL is WAY faster using a binary field instead of a varchar so I encode my INDEX in MD5 using BINARY(16)
to store the data.
This morning, I've started to use CRC32 for the first time and I saw that CRC32 can be output as an hexadecimal string using 8 characters.
My question : If I use a CRC32 instead of a MD5, it will be faster. However, when CRC32 is ran over let's say 2 000 000 unique value, the result will be unique or maybe sometime I'll have twice the same string for two differents string ? I ask that because the result is only 8 characters (32b) long instead of 32(128b) like the MD5.
Thanks.