Question

my SQL query
SELECT keyword
FROM table
WHERE MATCH (keyword)
AGAINST ('eco*' IN BOOLEAN MODE);
matches cells with these words:economy,ecology,echoscopy(why?),echo(why?) etc.

other SQL query
SELECT keyword
FROM table
WHERE MATCH (keyword)
AGAINST ('eci*' IN BOOLEAN MODE);
matches cell with the word:echidna.

Yet both queries don’t match word ectoplasm.

Why does echo, echoscopy match 'eco*' and echidna matches 'eci*'?

I see key element in this problem being letter combination “ch“.

Why does it work this way and how can I avoid this kind of matching?

Answer 1

The problem (feature?) was in the collation. “c” and “ch” were treated equal because of utf8_lithuanian_ci collation.

Edit:

Changing collation to utf8_unicode_ci fixes only certain issues.

The real solution is to use utf8_bin, which matches binary values of each character, meaning it’s:

• case sensitive
• diacritics sensitive