Stemming analyzer with language

[MattW]

I'm posting this in here as well, as the issue goes away when I change the "Change Analyzer" with ES.

Thread in RM support : http://xenforo.com/community/threads/tinhte-xentag.32231/page-27#post-509601



Basically, with the Tag mod, certain words aren't being picked up on the tag page when using Stemming analyzer with language: English






Here is the debug from those pages (no content)


Found content


I've changed it to Standard analyzer and the threads are now found.

The words are found in all the other searches on the site, apart from that one page, so I'm wondering if there is something I can look at with ES or if it's the tag mod?

Installed Java and ES using Slavik's guide on here.

[root@vps2 elasticsearch]#java -version
java version "1.7.0_17"
Java(TM) SE Runtime Environment (build 1.7.0_17-b02)
Java HotSpot(TM) 64-Bit Server VM (build 23.7-b01, mixed mode)

[root@vps2 elasticsearch]#less z22se.log
[2013-03-30 12:16:24,880][INFO ][node ] [The Grip] {0.90.0.Beta1}[16459]: initializing ...
[2013-03-30 12:16:24,888][INFO ][plugins ] [The Grip] loaded [], sites []
[2013-03-30 12:16:27,541][INFO ][node ] [The Grip] {0.90.0.Beta1}[16459]: initialized
[2013-03-30 12:16:27,541][INFO ][node ] [The Grip] {0.90.0.Beta1}[16459]: starting ...
[2013-03-30 12:16:27,667][INFO ][transport ] [The Grip] bound_address {inet[/0:0:0:0:0:0:0:0:9300]}, publish_address {inet[/95.154.207.109:9300]}
[2013-03-30 12:16:27,723][WARN ][discovery.zen.ping.multicast] [The Grip] failed to send multicast ping request: IOException[Operation not permitted]
[2013-03-30 12:16:29,230][WARN ][discovery.zen.ping.multicast] [The Grip] failed to send multicast ping request: IOException[Operation not permitted]
[2013-03-30 12:16:30,737][INFO ][cluster.service ] [The Grip] new_master [The Grip][ltoE3MGzRYe3MYCVCVDmMQ][inet[/95.154.207.109:9300]], reason: zen-disco-join (elected_as_master)
[2013-03-30 12:16:30,768][INFO ][discovery ] [The Grip] z22se/ltoE3MGzRYe3MYCVCVDmMQ
[2013-03-30 12:16:30,804][INFO ][http ] [The Grip] bound_address {inet[/0:0:0:0:0:0:0:0:9200]}, publish_address {inet[/95.154.207.109:9200]}
[2013-03-30 12:16:30,804][INFO ][node ] [The Grip] {0.90.0.Beta1}[16459]: started
[2013-03-30 12:16:32,040][INFO ][gateway ] [The Grip] recovered [1] indices into cluster_state

 

[MattW]

I've just installed a version of 1.1.3 and XFES 1.0.0, and I get the same results when using Stemming analyzer

It's weird though, because for an example, if I use a list of words (these are all models of cars)

vectra (works)
vectra a (doesn't work)
vectra b (works)
vectra c (doesn't work)
vectra d (works)
vectra e (works)
vectra f (works)

another list

astra (works)
astra a (works)
astra b (works)
astra c (works)
astra d (doesn't work)
astra e (works)
astra f (works)
astra g (works)
astra h (works)
astra j (doesn't work)

However, post astra j in a thread, and ES find it no problem

 

[Mike]

I'd guess it's down to how the tag add-on is firing its searches off, at least because the regular search works. You may need to dive into the code or contact the author.

 

[MattW]

Thanks Mike, I've posted the same details in the authors thread.

 

[MattW]

Managed to fix this (I think)!

I've no idea why or what was actually happening, but I've deleted the index via the ACP, done a full refresh, and it's now working again??

EDIT: Scrap that, when you delete the index, it defaults back to standard analyzer which is why it was working