I have this situation
SELECT field FROM index WHERE MATCH(‘“Chauffeur C”’);
1.35s
SELECT field FROM index WHERE MATCH(‘“Chauffeur”’);
30ms
Relevant index config:
max_filter_values = 100000
threads = 8
access_plain_attrs = mlock
access_blob_attrs = mlock
access_doclists = mlock
access_hitlists = mlock
max_threads_per_query = 1
not_terms_only_allowed = 1
Relevant serch config:
html_strip = 1
min_infix_len = 5
min_prefix_len = 1
expand_keywords = 1
index_exact_words = 1
Is this expected behaviour? Is a specific setting causes this?
I would expect it to be superfast.
tomat
July 9, 2025, 7:06am
2
if you check show meta
you would see that Chauffeur C
searches for *Chauffeur* *C*
and *C*
matches alot of different words that slow down the matching
The query is in quotes. I would expect it should only search for exactly this string:
Like this:
*Chauffeur C*
Is this not the case? Can I change the behaviour somehow?
tomat
July 9, 2025, 7:22am
4
the manual said about expand_keywords
Creating a table > NLP and tokenization > Wildcard searching settings | Manticore Search Manual and show the example of how it expands the terms
For instance,running
will become (running | *running* | =running)
It works inside all operators include phrase.
If you really need it enabled per index you could turn it off for some query wit hthe query options Searching > Options | Manticore Search Manual