Open munggruel opened 8 years ago
Yes, we will.
import full text search column error
@h1100e Full text search is still on the way.
Does it make sense to support this ? Since when you go lots of distributed data, you may want to use a specialized tool like solr,elasticsearch etc ?
Solr or elastic-search would be a better choice. Full-text search is on our roadmap. This could make it more convenient for accessing your data in TiDB in a different way. But it is not a high priority feature.
Just a quick update: TiDB 3.0 will not yet support FULLTEXT
, but we plan to make it a little clearer that it is not supported: https://github.com/pingcap/tidb/pull/9821
Support FULLTEXT search support 中文分词
@morgo @shenli Any idea how long it might take to implement FULLTEXT
? Popular open-source apps like Mattermost need it and our project needs some sort of full text search as well. Just asking before we decide for a strategy. Sending you a complete pull request doesn't seem feasible for now.
@lastzero Thanks for your interest in TiDB. The FULLTEXT
index feature is not included in TiDB 4.0 (which will be released in 2020 Q1) plan because of the limited resource. We will consider it in the next year. Community contributions are always welcomed.
@shenli Would you consider adding support to Trigger, so that a Elasticsearch-bridge that maps the updates to the Elasticsearch, like zombodb, can be implemented?
Elasticsearch has very good relevance algorithm and many plugins for different languages, and it is also distributed.
@shenli Will FULLTEXT
be supported in 5.0?
@wenfengwang sorry, it is not on the current 5.0 plan. It is a big feature: please thumbs up so we can track relative interest. Thanks!
The official website also mentions Spatial functions
FULLTEXT/SPATIAL functions and indexes #1793
but this ticket has nothing to do with spatial search.
It would be nice to see this feature placed into a milestone.
@munggruel I realize this is very old but I'm re-visiting this idea. Can you provide some information on the use case for this? It's easy to assume what the use cases are but it helps to know more specifically.
If it's not longer relevant to you, any information about the past use case would be helpful too, if you have the time. You're welcome to find me on the community Slack as well.
I'm obviously not @munggruel, but I really need this.
my use case: users record free-form text notes. months or years later they need to search those notes to find out when they mentioned this or that.
currently we have to read through every single one of the notes to find those that contain the substrings that are being searched for. this is obviously very slow.
if TiDB could index ngrams in text, it would be much more efficient to search through free-form text.
if TiDB could index ngrams in text, it would be much more efficient to search through free-form text.
Like this? https://dev.mysql.com/doc/refman/8.0/en/fulltext-search-ngram.html
I would consider the WITH PARSER
part to be the next step after "basic" fulltext index support.
@kaeverens out of curiosity, do you have an in-house parser for the text or would a databases's default parser be good enough?
in my case, I would be perfectly happy with the default parser. I don't need anything fancy - just a reasonably fast general search.
wondering if this is still happening or skipped.
This would be a great feature to have, especially since planetscale killed its hobby tier.
Will you have fulltext search support plan in future