Problem with english analyzer

Chris_Berkhout · November 28, 2011, 10:38am

Hi All,

I think there's been a problem with the english analyzer since I upgraded
to v0.18.4.

It looks like the stemming is being done for query strings, but not when
indexing data.

Here is a gist recreation:

gist.github.com

https://gist.github.com/chrisberkhout/76a009affd2df4b91ccd

gistfile1.txt

# Using the english analyzer...

curl -X DELETE "http://localhost:9200/test"
curl -X POST "http://localhost:9200/test" -d '{
    "settings": {
        "number_of_shards": 1,
        "analysis": {
            "analyzer": {
                "default": {
                    "type": "english"

This file has been truncated. show original

Have I got it right?

Cheers,
Chris

kimchy · November 28, 2011, 11:28am

This might relate to a problem with the _all field and setting a default
analyzer. It has been fixed in 0.18 branch (and upcoming 0.18.5), can you
give it a go?

On Mon, Nov 28, 2011 at 12:38 PM, Chris Berkhout chrisberkhout@gmail.comwrote:

Hi All,

I think there's been a problem with the english analyzer since I upgraded
to v0.18.4.

It looks like the stemming is being done for query strings, but not when
indexing data.

Here is a gist recreation:
problem with english analyzer in ES 0.18.4? · GitHub

Have I got it right?

Cheers,
Chris

kimchy · November 28, 2011, 12:59pm

Just double checked and it seems to be fixed (based on your gist) in 0.18,
and will be part of 0.18.5.

On Mon, Nov 28, 2011 at 1:28 PM, Shay Banon kimchy@gmail.com wrote:

This might relate to a problem with the _all field and setting a default
analyzer. It has been fixed in 0.18 branch (and upcoming 0.18.5), can you
give it a go?

On Mon, Nov 28, 2011 at 12:38 PM, Chris Berkhout chrisberkhout@gmail.comwrote:

Hi All,

I think there's been a problem with the english analyzer since I upgraded
to v0.18.4.

It looks like the stemming is being done for query strings, but not when
indexing data.

Here is a gist recreation:
problem with english analyzer in ES 0.18.4? · GitHub

Have I got it right?

Cheers,
Chris

Chris_Berkhout · November 29, 2011, 2:55am

Thanks Shay,

Just built the 0.18 branch (at commit 194106), and it worked.

Other problems I'd noticed with queries sometimes (not always) failing
with 500 also disappeared.

Cheers,
Chris

On Mon, Nov 28, 2011 at 8:59 PM, Shay Banon kimchy@gmail.com wrote:

Just double checked and it seems to be fixed (based on your gist) in 0.18,
and will be part of 0.18.5.

On Mon, Nov 28, 2011 at 1:28 PM, Shay Banon kimchy@gmail.com wrote:

This might relate to a problem with the _all field and setting a default
analyzer. It has been fixed in 0.18 branch (and upcoming 0.18.5), can you
give it a go?

On Mon, Nov 28, 2011 at 12:38 PM, Chris Berkhout chrisberkhout@gmail.com
wrote:

Hi All,
I think there's been a problem with the english analyzer since I upgraded
to v0.18.4.
It looks like the stemming is being done for query strings, but not when
indexing data.
Here is a gist recreation:
problem with english analyzer in ES 0.18.4? · GitHub
Have I got it right?
Cheers,
Chris