Re: Shield doesn't support third party plugins

[Tim S]

Thanks Uri, that helps, it makes a lot of sense.

On Wednesday, January 28, 2015 at 10:50:17 AM UTC, uboness wrote:
>
> Tim,
>
> We're in the process of clarifying this in the docs (agreed that the 
> current description is not really clear). Let me try to clarify it a bit 
> here...
>
> When it comes to third party plugins, we have no control over the plugin 
> code. The plugin infrastructure is extremely flexible in terms of what can 
> be extended in elasticsearch, from adding analyzers to adding new internal 
> actions and rest endpoints. While the former will have no impact on 
> security, the latter might have a significant impact and potentially 
> completely bypass the security checks in the system. For this reason, from 
> a company perspective, we can't really support plugins that are not under 
> our control (note that a lot of these plugins are developed internally in 
> companies and are not open source such that we can even review the code).
>
> As far as "won't work" is concerned, it obviously depends on what the 
> plugin is doing. A lot of plugins will work just fine (e.g. adding 
> additional analyzers), but others may experience unexpected behaviour when 
> developed without Shield security concerns in mind.
>
> I hope this clarifies it a bit. As mentioned above, we will fix the docs 
> with better explanation about it.
>
>
> On Wednesday, January 28, 2015 at 10:27:20 AM UTC+1, Tim S wrote:
>>
>> http://www.elasticsearch.org/guide/en/shield/current/limitations.html 
>> says that "Third-party plugins are not supported on clusters with the 
>> Shield security plugin installed."
>>
>> Can someone clarify the difference between "not supported" and "won't 
>> work" in this case please?
>>
>> If I have a plugin that is critical to the way I use elasticsearch (e.g. 
>> a plugin that adds a custom analyzer), is that page saying that 
>> Elasticsearch.com will not support an installation containing both shield 
>> and this analysis plugin? So that just means that anyone using third party 
>> plugins cannot use Shield at all? Is there any plan to change that?
>>
>> Thanks.
>>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/aa4dbf86-709d-4399-8496-6a94a358610e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Shield doesn't support third party plugins

[Tim S]

http://www.elasticsearch.org/guide/en/shield/current/limitations.html says 
that "Third-party plugins are not supported on clusters with the Shield 
security plugin installed."

Can someone clarify the difference between "not supported" and "won't work" 
in this case please?

If I have a plugin that is critical to the way I use elasticsearch (e.g. a 
plugin that adds a custom analyzer), is that page saying that 
Elasticsearch.com will not support an installation containing both shield 
and this analysis plugin? So that just means that anyone using third party 
plugins cannot use Shield at all? Is there any plan to change that?

Thanks.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/2d838ef0-209b-4475-8e0f-22734fff0472%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Which versions of elasticsearch will shield be compatible with?

[Tim S]

Hi,

Does anyone know which versions of elasticsearch shield can be used with? 
Will it just be the latest version at the time of release (1.4.2 currently) 
or will it work with older versions?

Thanks,

Tim.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/4e9e2cda-fed0-435a-8765-a8288b6b6335%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Comparison of int/long and float/double types

[Tim S]

So let's assume the cardinality is the same. Let's assume I have no text, I 
only index numeric fields.

If I've got a range of data, that would all fit within the bounds of an 
integer, is there any reason not to index it as a long? Are there any down 
sides? It sounds like you're saying that there isn't?



On Friday, October 24, 2014 2:55:24 PM UTC+1, Jörg Prante wrote:
>
> Lucene only knows how to index text strings. For numeric types, they are 
> stored as tries. Tries work on variable length. So only the API is 
> different to convert integer or long to tries. Tries are the basis for 
> numeric range searches.
>
> It is a myth that long take more disk space than ints in an inverted index 
> like Lucene. Both long and integer (numeric types) take a bit more space 
> than text strings, but for large indices, this does not add up at all, it 
> is in the noise.
>
> For field caches/filters, and doc values, the difference of integer and 
> long is more important. But there are other aspects like field cardinality 
> which determine the overall storage volume required.
>
> Jörg
>
> On Fri, Oct 24, 2014 at 3:22 PM, Tim S > 
> wrote:
>
>> I get the impression that using the 'long' type instead of 'integer' 
>> would use more disk space and degrade search performance (similary for 
>> double instead of float), but there's nothing in the documentation to back 
>> this impression up.
>>
>> There must be an advantage to using integer (if you can) because 
>> otherwise it wouldn't exist. It just doesn't say what the advantage is.
>>
>> Can someone confirm? Even better does anyone have any stats on what 
>> difference it would make?
>>
>> Thanks.
>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "elasticsearch" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to elasticsearc...@googlegroups.com .
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/elasticsearch/ff73f7bc-0578-44ea-803f-39fe723b5764%40googlegroups.com
>>  
>> <https://groups.google.com/d/msgid/elasticsearch/ff73f7bc-0578-44ea-803f-39fe723b5764%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/eff095af-0544-4feb-b09f-fa5c0158a4f5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Comparison of int/long and float/double types

[Tim S]

I get the impression that using the 'long' type instead of 'integer' would 
use more disk space and degrade search performance (similary for double 
instead of float), but there's nothing in the documentation to back this 
impression up.

There must be an advantage to using integer (if you can) because otherwise 
it wouldn't exist. It just doesn't say what the advantage is.

Can someone confirm? Even better does anyone have any stats on what 
difference it would make?

Thanks.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/ff73f7bc-0578-44ea-803f-39fe723b5764%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: final request when scroll-scanning has 0 successful shards

[Tim S]

Yes, it makes sense in this case, it's just confusing because it happens 
differently in other situations - when doing a normal scroll (not 
scanning), shards.successful is non-zero even when you've reached the point 
where there's no more results (and even if you keep going). And if you do a 
search returning no results, shards.successful is also non-zero.

So, should I ever (not just when scrolling) consider shards.successful=0 to 
indicate a failure? Or should I ignore it and only look at the 
shards.failed?

On Monday, August 4, 2014 12:27:22 PM UTC+1, Clinton Gormley wrote:
>
> This is correct. On the last request, no hits are returned because all 
> shards have already been drained of results.  If you look at shards.total 
> and shards.failed, you'll see they are also 0
>
> clint
>
>
> On 4 August 2014 12:54, Tim S > wrote:
>
>> When scroll-scanning 
>> <http://www.elasticsearch.org/guide/en/elasticsearch/reference/1.x/search-request-scroll.html#scroll-scan>,
>>  
>> I keep scrolling until I get a result that returns 0 docs, which is what 
>> the docs seem to suggest that I should do.
>>
>> But the final request (the one that returns 0 docs), always has 
>> _shards.successful=0. Normally, I’d consider this to indicate a failure 
>> (no shards responded). If doing a normal (non-scan) scroll, 
>> _shards.successful=1 (I’ve got 1 shard in this case), even when I’ve got 
>> to the point where there’s no more docs to scroll.
>>
>> Is this a bug in the scanning? Or am I wrong to consider the 
>> _shards.successful=0 to indicate a failure?
>>
>> Tested on version 1.3.1.
>>  
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "elasticsearch" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to elasticsearc...@googlegroups.com .
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/elasticsearch/3998b393-476d-4fc4-a809-a15da19bdf8d%40googlegroups.com
>>  
>> <https://groups.google.com/d/msgid/elasticsearch/3998b393-476d-4fc4-a809-a15da19bdf8d%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/03dc5c45-01e3-4d1e-a23c-830a26ef4f3b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

final request when scroll-scanning has 0 successful shards

[Tim S]

 

When scroll-scanning 
,
 
I keep scrolling until I get a result that returns 0 docs, which is what 
the docs seem to suggest that I should do.

But the final request (the one that returns 0 docs), always has 
_shards.successful=0. Normally, I’d consider this to indicate a failure (no 
shards responded). If doing a normal (non-scan) scroll, _shards.successful=1 
(I’ve got 1 shard in this case), even when I’ve got to the point where 
there’s no more docs to scroll.

Is this a bug in the scanning? Or am I wrong to consider the 
_shards.successful=0 to indicate a failure?

Tested on version 1.3.1.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/3998b393-476d-4fc4-a809-a15da19bdf8d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

How can I find out which date/number field matched?

[Tim S]

If the user does a complex search (lots of ORs and ANDs in their query), 
it's useful to be able to show them results, highlighting the bits in each 
one that matched. The highlighting feature does this fine for string 
fields, but it does not appear to work for dates and numbers.

Is there any way of finding out which fields matched?

Thanks,

Tim.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/e0b19706-1c43-4ddd-89bc-5c9e06c3508d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Sense on github abandoned?

[Tim S]

@kimchy the whole reason for me asking these questions is that sometimes a 
customer is using elasticsearch but they don't (yet) have a support 
contract, but don't consider themselves "in development" either, and thus 
wouldn't allow me to use Marvel. Yes, there are other tools for poking 
around, but sense is invaluable for constructing complicated queries etc 
quickly. In this situation they wouldn't let me install a chrome plugin 
either, but sense works nicely as an elasticsearch plugin too.

So, if sense (the abandoned version on github) had some kind of permissive 
licence, I could turn up on customer site and use sense to poke around.
Ideally, it would have a licence like AL2 which would allow me to modify it 
if necessary.

I realise that you don't want updates pushed back to the version of sense 
on github because those changes are helping you to make money from Marvel, 
I understand that. But if the abandoned version of sense did have an 
appropriate licence, it would allow us to use the current version - it's 
still useful even if it's not kept up to date. I might even be tempted to 
try and keep it up to date in my spare time. But clearly I can't do this 
unless it has a licence that allows me to do it.

Glad to see I'm not the only person thinking along these lines.


On Tuesday, April 1, 2014 11:15:07 AM UTC+1, Jörg Prante wrote:
>
> +1 for Sense standalone packaging
> +1 for Sense in Chrome Web Store
>
> Sense is used here all the time, it's essential.
>
> I have also forked the code in case Sense goes away, hoping for a FOSS 
> license.
>
> Not that I'm fluid in writing browser plugins, but if I find time, I am 
> not afraid of the learning curve.
>
> Jörg
>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/837794c8-1a0a-411f-a29c-852133d6fbc2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Sense on github abandoned?

[Tim S]

Yeah, I cloned it locally in case it disappears and it seems to work 
perfectly well standalone or as an elasticsearch plugin - IMO it doesn't 
really need to be a chrome extension. But I guess its usefulness if limited 
if dev has stopped and the auto-complete isn't kept up-to-date with the 
elasticsearch api.

Also if I recommended my clients used it (which I would, because it's one 
of the better rest front ends for ES I've come across because of the 
auto-complete) then they'd want to know whether they were actually allowed 
to use it i.e. what licence it is.


On Friday, March 28, 2014 2:50:04 PM UTC, Ivan Brusic wrote:
>
> I just forked it i case it goes away. I haven't built a Chrome extension 
> in years. Let me re-figure out how to do it and update my fork with 
> build/local installation instructions.
>
> The lack of a license might be problematic since default copyright 
> provisions apply: http://choosealicense.com/no-license/
>
> -- 
> Ivan
>
>
> On Fri, Mar 28, 2014 at 1:37 AM, Tim S >wrote:
>
>> I notice that https://github.com/bleskes/sense has a message saying "The 
>> development of Sense has moved into Elasticsearch Marvel".
>>
>> Does this mean that no further development will happen on github? I.e. if 
>> the Marvel team find bugs in sense will the fixes be pushed to the sense on 
>> github, and if I create a pull request on github will my fix find its way 
>> into the version of sense included in Marvel?
>>
>> Regardless of the answer to the above, does the code on github have any 
>> kind of licence? Even without Marvel, sense is still a useful tool.
>>
>> Thanks.
>>  
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "elasticsearch" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to elasticsearc...@googlegroups.com .
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/elasticsearch/381ca707-f29c-4b81-8e5b-957c1c02fe0a%40googlegroups.com<https://groups.google.com/d/msgid/elasticsearch/381ca707-f29c-4b81-8e5b-957c1c02fe0a%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/a8c17fc2-0f74-4161-992f-0a3e77e7153b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Sense on github abandoned?

[Tim S]

I notice that https://github.com/bleskes/sense has a message saying "The 
development of Sense has moved into Elasticsearch Marvel".

Does this mean that no further development will happen on github? I.e. if 
the Marvel team find bugs in sense will the fixes be pushed to the sense on 
github, and if I create a pull request on github will my fix find its way 
into the version of sense included in Marvel?

Regardless of the answer to the above, does the code on github have any 
kind of licence? Even without Marvel, sense is still a useful tool.

Thanks.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/381ca707-f29c-4b81-8e5b-957c1c02fe0a%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Bug in Java PutMapping API?

[Tim S]

@Andre might be a stupid question, but why are you trying to set a format 
for a string field and set 'index' on a long field? Afaics from the guide 
(http://www.elasticsearch.org/guide/en/elasticsearch/reference/1.x/mapping-core-types.html)
 
neither of these things are supported, so is it that surprising that 
elasticsearch drops the invalid attributes?

On Wednesday, February 26, 2014 11:51:32 PM UTC, Andre Encarnacao wrote:
>
> I am trying to use the Java API (v1.0.0) to create a custom mapping for my 
> index but have run into a problem. More specifically, some of my mapping 
> fields (for example, "format" and "index") are not being stored as part of 
> the mapping in Elastic Search. In fact, the only field that is being stored 
> is the "type" field. However, if I use the CURL API, everything works as 
> expected and all fields are stored properly in my mapping. Has anyone seen 
> this problem when using the Java API? Is this a bug? I put together a very 
> simple example that demonstrates my bug below:
>
>
> --- SOURCE CODE ---
>
> Client client = new TransportClient().addTransportAddress(new 
> InetSocketTransportAddress(_hostname,_port));
>
> XContentBuilder mapping = XContentFactory.jsonBuilder()
> .startObject().startObject("type_name")
> .startObject("properties")
> .startObject("ATTRIBUTE1")
> .field("type","string")
> .field("format","dateOptionalTime")
> .endObject()
> .startObject("ATTRIBUTE2")
> .field("type","long")
> .field("index","not_analyzed")
> .endObject()
> .endObject()
> .endObject()
> .endObject();
> 
>
> client.admin().indices().preparePutMapping("index_name").setType("type_name").setSource(mapping).execute().actionGet();
> 
>
> --- OUTPUT ---
>
> Below is the mapping created in ES after running the code snippet above 
> (from http://localhost:9200/index_name/_mapping?pretty). Notice that only 
> the "type" fields were stored in the mapping.
> {
>   "index_name" : {
> "mappings" : {
>   "type_name" : {
> "properties" : {
>   "ATTRIBUTE1" : {
> "type" : "string"
>   },
>   "ATTRIBUTE2" : {
> "type" : "long"
>   }
> }
>   }
> }
>   }
> }
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/fae41878-ce9d-41ea-9f84-24acd0657095%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

geo_point in nested has the 'path' attribute added to it

[Tim S]

I was writing some tests to check if my mappings were being deployed 
correctly, and came across this: if you have a geo_point field inside a 
nested object, it will inherit the 'path' attribute from the nested object.

I.e. if you create an index like this:

curl -XPOST 'localhost:9200/test' -d '{
"mappings" : {
"type1" : {
"properties" : {
"geoInOuter" : {
"type" : "geo_point"
},
"obj1" : {
"type" : "nested",
"path" : "just_name",
"properties" : {
"geoInNested" : {
"type" : "geo_point"
}
}
}
}
}
}
}'

and then ask for the mapping back, like this
curl -XGET 'localhost:9200/test/_mapping'

the response will be this:

{
"test" : {
"mappings" : {
"type1" : {
"properties" : {
"geoInOuter" : {
"type" : "geo_point"
},
"obj1" : {
"type" : "nested",
"path" : "just_name",
"properties" : {
"geoInNested" : {
"type" : "geo_point",
"path" : "just_name"
}
}
}
}
}
}
}
}

Notice the extra 'path' attribute on the 'geoInNested' field.

Is this deliberate? In which case what affect does the path attribute have 
on a geo point field? Or is this a bug?

Thanks.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/d60a0c7c-afb0-42f7-bfc0-19577189e34a%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Re: Any possibility of permalinks to documentation for each version?

[Tim S]

"It doesn't seem like they support minor versions." - this is exactly what 
I'm asking "them" to do...

It doesn't have to be linked to from the main page, but a stable link that 
I know will still be there a year or two in the future will be very useful.

On Monday, January 27, 2014 7:54:47 PM UTC, Ivan Brusic wrote:
>
> Here are the current branches: 
> http://www.elasticsearch.org/guide/en/elasticsearch/reference/index.html
>
>
> On Mon, Jan 27, 2014 at 11:53 AM, Ivan Brusic 
> > wrote:
>
>> The documentation is now versioned. For example (random page)
>>
>> 0.90:  
>> http://www.elasticsearch.org/guide/en/elasticsearch/reference/0.90/search-search.html
>> master: 
>> http://www.elasticsearch.org/guide/en/elasticsearch/reference/master/search-search.html
>> current: 
>> http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-search.html
>>
>> I am guessing a 1.0 version will be built once it is released. It doesn't 
>> seem like they support minor versions. Not sure what "current" signifies, 
>> but you can probably figure it out by reading the documentation building 
>> script source: https://github.com/elasticsearch/docs
>>  
>> -- 
>> Ivan
>>
>>
>> On Mon, Jan 27, 2014 at 1:27 AM, Tim S > >wrote:
>>
>>> I notice the docs for version 1.0.0 are in a "1.x" branch. When 1.1 is 
>>> released, I'm assuming the 1.x branch docs will be updated to 1.1?
>>>
>>> Is there any possibility of creating permanent links to each version? 
>>> (like lucene do, e.g http://lucene.apache.org/core/3_6_1/). The reason 
>>> is simply that if I build a solution for a customer that requires a 
>>> specific version of elasticsearch, then my documentation will probably end 
>>> up referencing various bits of the elasticsearch documentation. Which is 
>>> tricky if the documentation I'm referencing is changing.
>>>
>>> So, I was wondering if it would be possible to create links for each 
>>> release version, so instead of referencing 
>>> http://www.elasticsearch.org/guide/en/elasticsearch/reference/1.x/search.htmlI
>>>  can reference 
>>> http://www.elasticsearch.org/guide/en/elasticsearch/reference/1.0.0.Release/search.htmlor
>>>  similar, and then I can be happy that the documentation I'm linking to 
>>> is the exact documentation for the version that my customer uses.
>>>
>>> Thanks,
>>>
>>> Tim.
>>>
>>> -- 
>>> You received this message because you are subscribed to the Google 
>>> Groups "elasticsearch" group.
>>> To unsubscribe from this group and stop receiving emails from it, send 
>>> an email to elasticsearc...@googlegroups.com .
>>> To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/elasticsearch/dd467b38-da9b-4f4e-9ebe-f39adb403f2f%40googlegroups.com
>>> .
>>> For more options, visit https://groups.google.com/groups/opt_out.
>>>
>>
>>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/88e08dda-b9c5-4bdd-b6e3-73bba41ca503%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Any possibility of permalinks to documentation for each version?

[Tim S]

I notice the docs for version 1.0.0 are in a "1.x" branch. When 1.1 is 
released, I'm assuming the 1.x branch docs will be updated to 1.1?

Is there any possibility of creating permanent links to each version? (like 
lucene do, e.g http://lucene.apache.org/core/3_6_1/). The reason is simply 
that if I build a solution for a customer that requires a specific version 
of elasticsearch, then my documentation will probably end up referencing 
various bits of the elasticsearch documentation. Which is tricky if the 
documentation I'm referencing is changing.

So, I was wondering if it would be possible to create links for each 
release version, so instead of referencing 
http://www.elasticsearch.org/guide/en/elasticsearch/reference/1.x/search.htmlI 
can reference 
http://www.elasticsearch.org/guide/en/elasticsearch/reference/1.0.0.Release/search.htmlor
 similar, and then I can be happy that the documentation I'm linking to 
is the exact documentation for the version that my customer uses.

Thanks,

Tim.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/dd467b38-da9b-4f4e-9ebe-f39adb403f2f%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Re: How can I merge the results of two aggregations?

[Tim S]

Sorry, maybe I'm missing something. Afaics the OR filter would operate on 
filters or queries, not on aggregations. Can you give me an example of how 
I'd use this to merge the result of the two aggregations?

Thanks.

On Friday, December 20, 2013 2:54:14 AM UTC, kidkid wrote:
>
> Hi sorry, it would be my mistake:
>
> Could you take a look at OrFilter: 
> http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-or-filter.html
>
> In your case I think you could use match all query & use OrFilter & let ES 
> merge the result.
>
>
>
>
>
> On Thursday, December 19, 2013 2:29:06 AM UTC-8, Tim S wrote:
>>
>> How can I use bool query on the result of an aggregation? I want both 
>> aggregations to independently facet on the whole index, then merge the 
>> results. I can see how I would use a bool query to limit the set of docs 
>> I'm aggregating on, but I can't see how I would use it to merge the results 
>> of two aggregations?
>>
>> Thanks,
>>
>> Tim.
>>
>> On Wednesday, December 18, 2013 4:40:37 PM UTC, kidkid wrote:
>>>
>>> You could use bool query and let ElasticSearch do the rest.
>>>
>>>
>>> http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-bool-query.html
>>>
>>> On Wednesday, December 18, 2013 3:58:07 PM UTC+7, Tim S wrote:
>>>>
>>>> In the example below, I ask elasticsearch for two aggregations (I've 
>>>> simplified it, it's actually got some nested aggregations in there).
>>>>
>>>> {
>>>> "aggregations": {
>>>> "agg1": {
>>>> "terms": {
>>>> "field": "forname"
>>>> }
>>>> },
>>>> "agg2": {
>>>> "terms": {
>>>> "field": "surname"
>>>> }
>>>> }
>>>> }
>>>> }
>>>>
>>>> What I get back is two sets of results, i.e. 
>>>>
>>>> {
>>>> "aggregations" : {
>>>> "agg1" : {
>>>> "buckets" : [ {
>>>> "key" : "john",
>>>> "doc_count" : 1
>>>> }, {
>>>> "key" : "bob",
>>>> "doc_count" : 4
>>>> } ]
>>>> },
>>>> "agg2" : {
>>>> "buckets" : [ {
>>>> "key" : "smith",
>>>> "doc_count" : 3
>>>> }, {
>>>> "key" : "jones",
>>>> "doc_count" : 2
>>>> } ]
>>>> }
>>>> }
>>>> }
>>>>
>>>> What I'd like to get back is one set of results. I.e. a list of terms 
>>>> appearing in either of the fields, with the counts summed across both, e.g.
>>>>
>>>> {
>>>> "aggregations" : {
>>>> "agg1 OR agg2" : {
>>>> "buckets" : [ {
>>>> "key" : "john",
>>>> "doc_count" : 1
>>>> }, {
>>>> "key" : "bob",
>>>> "doc_count" : 4
>>>> }, {
>>>> "key" : "smith",
>>>> "doc_count" : 3
>>>> }, {
>>>> "key" : "jones",
>>>> "doc_count" : 2
>>>> } ]
>>>> }
>>>> }
>>>> }
>>>>
>>>> Is there any way of doing this? I could request them and merge them in 
>>>> my own code, but if there's a built in way then I'd rather use that.
>>>>
>>>> Thanks,
>>>>
>>>> Tim.
>>>>
>>>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/067e2d3e-2d7f-46e9-9ce6-0384a9a89b43%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Does the dfs search type work across indexes?

[Tim S]

If I use the search type "dfs_query_then_fetch", as I understand it the 
first step of the search is to go to each of the shards collecting the 
inverse document frequencies.

If I was searching across multiple indexes, would it collect the idfs from 
each of the indexes and combine them?

E.g. if I randomly split an index in half (e.g. "unsecured" and "secured"), 
if I did a search across both of those indexes, would the results (and 
their scores) be roughly similar to if I had all the documents in one index 
and just did the search on that?

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/f0cf0477-bb7e-4770-aedc-3e37e5d43414%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Re: How can I merge the results of two aggregations?

[Tim S]

How can I use bool query on the result of an aggregation? I want both 
aggregations to independently facet on the whole index, then merge the 
results. I can see how I would use a bool query to limit the set of docs 
I'm aggregating on, but I can't see how I would use it to merge the results 
of two aggregations?

Thanks,

Tim.

On Wednesday, December 18, 2013 4:40:37 PM UTC, kidkid wrote:
>
> You could use bool query and let ElasticSearch do the rest.
>
>
> http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-bool-query.html
>
> On Wednesday, December 18, 2013 3:58:07 PM UTC+7, Tim S wrote:
>>
>> In the example below, I ask elasticsearch for two aggregations (I've 
>> simplified it, it's actually got some nested aggregations in there).
>>
>> {
>> "aggregations": {
>> "agg1": {
>> "terms": {
>> "field": "forname"
>> }
>> },
>> "agg2": {
>> "terms": {
>> "field": "surname"
>> }
>> }
>> }
>> }
>>
>> What I get back is two sets of results, i.e. 
>>
>> {
>> "aggregations" : {
>> "agg1" : {
>> "buckets" : [ {
>> "key" : "john",
>> "doc_count" : 1
>> }, {
>> "key" : "bob",
>> "doc_count" : 4
>> } ]
>> },
>> "agg2" : {
>> "buckets" : [ {
>> "key" : "smith",
>> "doc_count" : 3
>> }, {
>> "key" : "jones",
>> "doc_count" : 2
>> } ]
>> }
>> }
>> }
>>
>> What I'd like to get back is one set of results. I.e. a list of terms 
>> appearing in either of the fields, with the counts summed across both, e.g.
>>
>> {
>> "aggregations" : {
>> "agg1 OR agg2" : {
>> "buckets" : [ {
>> "key" : "john",
>> "doc_count" : 1
>> }, {
>> "key" : "bob",
>> "doc_count" : 4
>> }, {
>> "key" : "smith",
>> "doc_count" : 3
>> }, {
>> "key" : "jones",
>> "doc_count" : 2
>> } ]
>> }
>> }
>> }
>>
>> Is there any way of doing this? I could request them and merge them in my 
>> own code, but if there's a built in way then I'd rather use that.
>>
>> Thanks,
>>
>> Tim.
>>
>

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/b9c82b3b-c8a3-4783-96d8-b4d07809189f%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

How can I merge the results of two aggregations?

[Tim S]

In the example below, I ask elasticsearch for two aggregations (I've 
simplified it, it's actually got some nested aggregations in there).

{
"aggregations": {
"agg1": {
"terms": {
"field": "forname"
}
},
"agg2": {
"terms": {
"field": "surname"
}
}
}
}

What I get back is two sets of results, i.e. 

{
"aggregations" : {
"agg1" : {
"buckets" : [ {
"key" : "john",
"doc_count" : 1
}, {
"key" : "bob",
"doc_count" : 4
} ]
},
"agg2" : {
"buckets" : [ {
"key" : "smith",
"doc_count" : 3
}, {
"key" : "jones",
"doc_count" : 2
} ]
}
}
}

What I'd like to get back is one set of results. I.e. a list of terms 
appearing in either of the fields, with the counts summed across both, e.g.

{
"aggregations" : {
"agg1 OR agg2" : {
"buckets" : [ {
"key" : "john",
"doc_count" : 1
}, {
"key" : "bob",
"doc_count" : 4
}, {
"key" : "smith",
"doc_count" : 3
}, {
"key" : "jones",
"doc_count" : 2
} ]
}
}
}

Is there any way of doing this? I could request them and merge them in my 
own code, but if there's a built in way then I'd rather use that.

Thanks,

Tim.

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to elasticsearch+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/731d5734-25ff-4d29-ad11-3a9f6a382244%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.