query with multiple hashtags very slow

[ahheckel]

I am running linkding in a docker container (following the instructions on this page) on a Raspberry4 under a current RaspianOS 32bit with ~500 bookmarks which works very well so far.
The response is quick when searching for a single hashtag, however it can be very long (up to a minute) when combining multiple hashtags in a query (#tag1 #tag2). This is particularly true when that query yields more than just a few results.
I was wonderung whether the database query can be optimized for speed.
Kind regards,
(PS great project).

[sissbruecker]

I tested this a bit and can confirm that query performance degrades when using multiple tags. However in my tests (also on a RPie 4) I would need to add a lot more than 2 tags and even then the query would resolve in a few seconds (which is still too slow IMO, but far better than what you are experiencing).

Specifically for the use-case described here - for 500 bookmarks and 2 tags in the query 1 minute would be extremely bad. I see two possible issues here:

• You might have a lot of tag assignments in general - if you could check how many rows you have in your bookmarks_bookmark_tags table in the database that might help solving this
• Or there is just some other reason why your installation performs to poorly. However if regular searches perform well that seems unlikey.

[ulixxe]

Hi,
I closed #141 and use this issue from now on to signal that I search for only 1 tag, and I have to restart the docker container after several minutes of 100% CPU utilization.
In my bookmarks_bookmark_tags table, I have about 2.5k rows.

[ulixxe]

I found that the search speed is ok if I comment the below lines 98-101 from bookmarks/queries.py. Can some optimization be done on those lines?

for tag_name in query['tag_names']:
    query_set = query_set.filter(
        bookmark__tags__name__iexact=tag_name
    )

[ahheckel]

ulixxe's suggestion solves the problem in my case, too.