Judging by the output, the genre field is analyzed using the default
analyzer. Others can help debug if you provide your mapping. It is best to
use the get mapping API [1] since it shows what you actually have instead
of what you supplied at index creation.
Depending on your use case, you might
The vectorization issue is not constrained to OpenJDK, and is still present
in 7u45: https://twitter.com/thetaph1/status/423523708708208640
On Tue, Mar 4, 2014 at 3:52 PM, InquiringMind brian.from...@gmail.comwrote:
Jörg,
Just to clarify: The links below point to OpenJDK, not to the Oracle
The zip download available on github is not what you want. The format
required for plugins is different from the source download found on github.
Since it appears that you do not have download access, as Roland mentioned,
your last option is to clone the project in git and build it yourself with
You can always disable term frequencies on a field to eliminate the td-idf
issue, but then scoring would be affected and perhaps be more detrimental
than the original problem.
The standard solution in Lucene is to use payloads, which is metadata
associated with a term in the index. The synonym
? Or can this be achieved with scripts in ES?
Thanks for your help,
Jake
On Monday, March 3, 2014 7:36:26 AM UTC-8, Ivan Brusic wrote:
You can always disable term frequencies on a field to eliminate the
td-idf issue, but then scoring would be affected and perhaps be more
detrimental than
I did some stress tests with different versions of the JDK using a test
cluster running on VMs (which only contains a smaller amount of data than
our production cluster), and I was never able to uncover any issues.
However, since I do not like getting called in the middle of the night to
fix a
I purchased the book when Packt was having a $5 ebook sale a couple of
months ago. Did not really need the book, but it was cheap and I wanted to
support the author who has posted on the mailing list in the past.
Overall a decent book, recommended for anyone getting started with
Elasticsearch. My
I have never tried or looked at the code, but off the top of my head
perhaps the DFS query type would work:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-request-search-type.html#dfs-query-then-fetch
Since the DFS query type calculates the TF/IDF values based on the
I do not use quotes at all. Simply:
node.name: ${HOSTNAME}
--
Ivan
On Tue, Feb 25, 2014 at 7:56 AM, InquiringMind brian.from...@gmail.comwrote:
I always start Elasticsearch from within my own wrapper script, es.sh.
Inside this wrapper script is the following incantation:
Luke? :)
On Tue, Feb 25, 2014 at 1:09 PM, Daniel Winterstein
daniel.winterst...@gmail.com wrote:
Dear Hariharan, Alex, Luke,
My apologies. You're quite right. The information is there -- I just
didn't read far enough down.
Thank you for your help persistence.
Best regards,
- Daniel
Ivan,
The DFS query then fetch worked very well!
Thank you!
Cheers,
Luiz Guilherme
On Tue, Feb 25, 2014 at 5:15 PM, Ivan Brusic i...@brusic.com wrote:
I have never tried or looked at the code, but off the top of my head
perhaps the DFS query type would work:
http://www.elasticsearch.org
David has the ideal solution, however just wanted to point out one key
difference in the get API is that it would look for a document in the
transaction log before it looks in the index. So if you execute a search
query after an insert, but before a refresh/flush, you will not see the
changes.
string based queries by filtering to a subset of documents first!
On Tuesday, February 18, 2014 3:19:50 PM UTC-6, Ivan Brusic wrote:
The documentation suddenly made me doubt if we I knew was wrong. :)
The default strategy for Elasticsearch's filtered query is a custom
random access one. For each
The boost field, like other special fields such as timestamp and id, is set
at the root level of a type, not as a property:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/mapping-boost-field.html
Perhaps I should have added this confusion as yet another reason why to
avoid
Hi Ivan (awesome name BTW),
Read my recent reply about filters:
https://groups.google.com/forum/?fromgroups#!topic/elasticsearch/xZGnyTI6lmo
Basically the filter in a filtered query is executed before the query. IMHO
that documentation is misleading. Use post filters (simply 'filter' before
1.0)
. The original description of the filtered query comes
directly from Lucene:
http://lucene.apache.org/core/4_6_0/core/org/apache/lucene/search/FilteredQuery.html
Cheers,
Ivan
On Tue, Feb 18, 2014 at 1:02 PM, Ivan Brusic i...@brusic.com wrote:
Hi Ivan (awesome name BTW),
Read my recent reply
Since your code is running as the same JVM as Elasticsearch, you have a few
more options. Depending on your use case, you can bypass the analysis API
and use the analyzers directly. Look at the test code in the analysis
package, in particular the AnalysisModuleTests class. The key is building
the
The change seemed simple, so I went ahead and implemented it:
https://github.com/elasticsearch/elasticsearch/pull/5148
The tricky part is getting the Elasticsearch team to notice. :)
--
Ivan
On Mon, Feb 17, 2014 at 2:13 PM, Ivan Brusic i...@brusic.com wrote:
Since your code is running
There is no automatic bucketing in Elasticsearch. I mimic the behavior with
an expensive process that uses many smaller fixed ranges which are reduced
into the number of buckets needed on the client side. Easily the slowest
part of my query. My goal was to wait for the facet refactor (which has
Here is a previous discussion on Rice/Sturges:
https://groups.google.com/forum/#!msg/elasticsearch/CAZhIHtB1UI/Exzd2_DanbAJ
Never did sit down and finally understand the paper Jörg linked. :) I
really should find the time to revisit the issue since my implementation is
costly.
Ivan
On Fri, Feb
I doubt this issue will ever be fixed since the limitation exists in
Lucene. All types belong to the same index and a field's data needs to be
uniform in Lucene's eyes. A document's type is used to indicate different
mappings for a document, but not different ways to segment the data types
in the
I'm glad I was able to steer you in the right direction. I flubbed a PR
recently since I have not used git consistently in the past few years, so I
am glad someone else can learn from my mistakes. Your PR seemed to have
gained some attention! :)
Ivan
On Tue, Feb 11, 2014 at 1:17 PM,
The documentation has not been correct for version 1.0 [1]. The method
should be now called setPostFilter. Better yet, you should look into
filtered queries [2].
[1] https://github.com/elasticsearch/elasticsearch/pull/4461
[2]
(FilterBuilders.termFilter(multi, test))
3) setQuery(QueryBuilders.filteredQuery(QueryBuilders.matchAllQuery(),
FilterBuilders.termFilter(multi, test)))
I'm guessing 3? But the matchAllQuery() makes it feel like that is wrong...
On Wed, Feb 12, 2014 at 9:25 PM, Ivan Brusic i...@brusic.com wrote
On Wed, Feb 12, 2014 at 1:58 PM, joergpra...@gmail.com
joergpra...@gmail.com wrote:
For my requirements, downtime of 15 min is acceptable.
I can only wish! I run an ecommerce site, so my requirement is no downtime.
Ever.
--
Ivan
--
You received this message because you are subscribed to
Great catch. Which Elasticsearch version and which JDK?
Thankfully my documents are uniform, so I have been able to skip isEmpty
checks.
--
Ivan
On Tue, Feb 11, 2014 at 7:52 AM, Nikolas Everett nik9...@gmail.com wrote:
Sorry to resurrect a dead thread, but I figured it out:
Very rookie problem. :)
The default (aka standard) analyzer uses a stopword filter and it is a
stopword. Try configuring your field with a custom analyzer which does not
use stopwords or a custom set of stopwords.
Cheers,
Ivan
On Tue, Feb 11, 2014 at 7:57 AM, felix.kof...@gameforge.de wrote:
are faster, can be cached and do not influence scoring.
Cheers,
Ivan
On Tue, Feb 11, 2014 at 9:00 AM, Ivan Brusic i...@brusic.com wrote:
Very rookie problem. :)
The default (aka standard) analyzer uses a stopword filter and it is a
stopword. Try configuring your field with a custom analyzer which
Try setting use_dis_max to false in your query.
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-query-string-query.html#_multi_field_2
Cheers,
Ivan
On Tue, Feb 11, 2014 at 1:15 AM, Alexander Ott
alexander.ott...@gmail.comwrote:
Hallo,
i have the same
Create a branch for your changes. Submit a PR from the branch and not
master. Make sure to update DistanceUnitTests.java as well. The trickiest
part is getting the Elasticsearch team to notice your PR. :) They must be
super busy with the 1.0 release.
Lots of tutorials online:
What is your current mapping? Use the GetMapping API.
The file field is an inner object, but you do not have one defined in your
mapping. Very likely you already have indexed a document with the file
field as another type.
--
Ivan
On Tue, Feb 11, 2014 at 7:12 AM, Stefan Sabolowitsch
That is your template. Use the Get Mapping API to find out what actually is
in effect.
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/indices-get-mapping.html
On Tue, Feb 11, 2014 at 12:17 PM, Stefan Sabolowitsch
sabolowitsc...@in-trier.de wrote:
Hi Ivan,
thanks for
Oops, obvious answer. :) I see questions about incorrect TFIDF scores and
my mind automatically goes to DFS scoring (which is actually about TF, not
IDF).
--
Ivan
On Tue, Feb 11, 2014 at 10:22 AM, Binh Ly b...@hibalo.com wrote:
Also be aware that the log should be a natural log, i.e. the
My understanding was that the scope of the JFK/Guava bug was limited and did
not affect Lucene/Solr. Elasticsearch uses Guava for collections and caching,
not sure about reflection.
--
Ivan
joergpra...@gmail.com wrote:
Do not use Java 7u51, Lucene bug still not fixed
is lucky...
Jörg
On Mon, Feb 10, 2014 at 4:00 PM, Ivan Brusic i...@brusic.com wrote:
My understanding was that the scope of the JFK/Guava bug was limited and
did not affect Lucene/Solr. Elasticsearch uses Guava for collections and
caching, not sure about reflection.
--
You received
Aircraft use nautical miles? You learn something new every day!
--
Ivan
On Mon, Feb 10, 2014 at 3:21 PM, InquiringMind brian.from...@gmail.comwrote:
Would it be reasonable to create an issue to request nautical miles (nm
as the abbreviation) for the DistanceUnit enumeration?
This would
The issues that plague 7u40 and 7u45 are related to Lucene and not directly
to Elasticsearch. Lucene is being continually being built with Java 7u51
and 7u60 and I do not think any major issues have been found yet. There are
several Lucene committers on this list, so maybe they will chime in (I
Relevancy is very context dependent. Only you have the knowledge to
determine if a result is relevant for the query or not.
Lucene/elasticsearch's default TFIDF algorithm should theoretically give
you the relevant results. Do the results change if you tweak your query
slightly?
Recall is even
implement RAFT or something
similar, I was there and heard that.
Otis
--
Performance Monitoring * Log Analytics * Search Analytics
Solr Elasticsearch Support * http://sematext.com/
On Wed, Feb 5, 2014 at 2:58 PM, Ivan Brusic i...@brusic.com wrote:
Otis, you can listen to the comment here
Interesting topic. Not elasticsearch specific, but nevertheless
interesting. One method to calculate relevancy given the precision and
recall of a query is by using the F1 score:
http://en.wikipedia.org/wiki/F1_score
F1 would be equal to 2 * (P * R) / (P + R), where P is the precision and R
is
Try using a ids filter combined with a not filter instead;
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-ids-filter.html
Filters are faster, cacheable (if desired) and will not influence scoring.
If you do not want the excluded ids to be added to the facet
So the versions of 192.168.100.210 and 192.168.100.90 are different?
Although elasticsearch is supposed to support cluster that differ between
minor versions, this scenario is no longer true with later versions of 0.90.
--
Ivan
On Mon, Feb 3, 2014 at 6:02 AM, Sambhav Sharma
?
On Fri, Jan 31, 2014 at 5:53 AM, Ivan Brusic i...@brusic.com wrote:
BTW, if you do not want to upgrade, you can also simply maintain two
fields. One to use as a document boost (the _boost) and another to use for
scripting/sorting/faceting.
--
Ivan
On Thu, Jan 30, 2014 at 4:00 PM, Ivan
Not sure what behavior you are observing since you did not add any results.
One thing however is you are essentially boosting everything twice. Once
with a document boost and then again with the custom score. Remove the
document boost.
--
Ivan
On Fri, Jan 31, 2014 at 12:49 PM, Mukul Gupta
I would add to flush the transaction log after you have indexed all your
content.
--
Ivan
On Fri, Jan 31, 2014 at 4:57 PM, joergpra...@gmail.com
joergpra...@gmail.com wrote:
Shutdown:
curl -XPOST node:9200/_shutdown
In the latest versions (1.0.0.RC1) ES shutdown chooses a strategy in
Query scores are not ordinal. Lucene discourages comparing scores from
different queries since the search context is important in scoring. Query
norms are supposed to help normalize scores across queries, but as you
discovered, that is not normally what happens. Your theory will probably
hold true
I fixed this bug a while ago and it should be fixed in versions 0.90.6 and
higher.
https://github.com/elasticsearch/elasticsearch/issues/3752
https://github.com/elasticsearch/elasticsearch/pull/2913
Ivan
On Thu, Jan 30, 2014 at 11:51 AM, Binh Ly b...@hibalo.com wrote:
Coder, I've verified
BTW, if you do not want to upgrade, you can also simply maintain two
fields. One to use as a document boost (the _boost) and another to use for
scripting/sorting/faceting.
--
Ivan
On Thu, Jan 30, 2014 at 4:00 PM, Ivan Brusic i...@brusic.com wrote:
I fixed this bug a while ago and it should
It appears that this value is not configurable. IMHO, it should be, so
perhaps you should open an issue on Github and see if the elasticsearch
team agrees. Better yet, submit a pull request. :)
--
Ivan
On Tue, Jan 28, 2014 at 1:08 PM, Ahaduzzaman Munna
ahaduzzaman.mu...@gmail.com wrote:
Marvel was just announced today (to me at least), and there already is a
question/issue? :) Let us know how it is working out.
I am assuming that the elasticsearch team has Marvel as a private repo,
which means you can't post issues.
--
Ivan
On Tue, Jan 28, 2014 at 7:45 AM, J. Schulz
)
.endObject()
.startObject(totalexp)
.field(store, yes)
.endObject()
.endObject()
.endObject()
.endObject()
.endObject();
---
Hiro
On Monday, 27 January 2014 23:50:41 UTC+5:30, Ivan Brusic wrote
Correct, create a plugin. Take a look at some of the existing analysis
plugins as a template:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/modules-plugins.html#analysis-plugins
--
Ivan
On Tue, Jan 28, 2014 at 8:48 AM, Guilhem Legal
guilhem.le...@geomatys.comwrote:
What you are probably looking for is field collapsing, which is not yet
supported in elasticsearch (it is planned). You can use a term facet to
retrieve the count for all terms and then do a separate query for each
unique term. In addition to the slowness of having to do multiple queries,
you
Jörg, what is the purpose of tweaking the IndexQueryParserService? I have a
few custom filters, and all I do is add them via
tokenFiltersBindings.processTokenFilter()
or analysisModule.addTokenFilter().
--
Ivan
On Tue, Jan 28, 2014 at 9:40 AM, joergpra...@gmail.com
joergpra...@gmail.com
It is only supported with Elasticsearch 0.90.8 and higher.
Yet another push for me to upgrade!
--
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
...@campaignmonitor.com
web: www.campaignmonitor.com
On 29 January 2014 09:37, Ivan Brusic i...@brusic.com wrote:
It is only supported with Elasticsearch 0.90.8 and higher.
Yet another push for me to upgrade!
--
You received this message because you are subscribed to the Google Groups
: ma...@campaignmonitor.com
web: www.campaignmonitor.com
On 29 January 2014 09:51, Ivan Brusic i...@brusic.com wrote:
From my log:
[2014-01-28 14:31:01,915][WARN ][marvel.agent ] Elasticsearch
version [0.90.2] is too old. Marvel is disabled (requires version 0.90.8 or
higher
Engineer
Campaign Monitor
email: ma...@campaignmonitor.com
web: www.campaignmonitor.com
On 29 January 2014 09:51, Ivan Brusic i...@brusic.com wrote:
From my log:
[2014-01-28 14:31:01,915][WARN ][marvel.agent ]
Elasticsearch version [0.90.2] is too old. Marvel is disabled
Marvel doesn't different much from elasticsearch plugins in that the code
now runs in the same JVM instead of a separate process. The event data is
pushed rather than pulled. It is great not having to re-invent the wheel,
but having monitoring outside of elasticsearch is not an issue. Great
For the third rule, you can omit index norms for a field which will prevent
length normalization. See [1]. The option is either called omit_norms
or norms.enabled depending on your version.
For the second rule, it is slightly more complicated. You can define your
own custom similarity [2] that
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 i...@brusic.com wrote:
The documentation is now versioned. For example (random page)
0.90:
http://www.elasticsearch.org/guide/en
://speakerdeck.com/benbjohnson/raft-the-understandable-distributed-consensus-protocol
Jörg
On Sat, Jan 25, 2014 at 7:26 PM, Ivan Brusic i...@brusic.com wrote:
Unfortunately the problem will still not be addressed in the upcoming 1.0
release. Judging by recent comments from the elasticsearch team
Unfortunately the problem will still not be addressed in the upcoming 1.0
release. Judging by recent comments from the elasticsearch team, they are
truly looking into the matter. The change will require changing the
underlying consensus algorithm to something like Paxos/Raft.
That said, the
Are you omitting norms by any chance on your fields?
--
Ivan
On Thu, Jan 23, 2014 at 12:32 PM, georgi.mat...@jobvector.com wrote:
What I want to do is do a fuzzy (like this) query across multiple fields
where matches in certain fields should result in a higher score than
matches in other
Can you establish a naming convention for the parent fields? You can
probably use index templates:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/indices-templates.html
--
Ivan
On Thu, Jan 23, 2014 at 1:55 PM, Olivier Rossel olivier.ros...@gmail.comwrote:
I mostly use
You can run zookeeper as its own cluster and not just a singleton. At this
point, you will now have an elasticsearch cluster and a zookeeper cluster!
Are you experiencing split brain issues? You might be better off determine
how to avoid situations that lead to split brain.
Cheers,
Ivan
On
The first thing to do is check if your limits are actually being persisted
and used. The elasticsearch site has a good writeup:
http://www.elasticsearch.org/tutorials/too-many-open-files/
Second, it might be possible that you are reaching the 128k limit. How many
shards per node do you have? Do
for this ?
Also, Can I do this sort in my mapping itself ?
Thanks
On Wed, Jan 22, 2014 at 3:22 AM, Ivan Brusic i...@brusic.com wrote:
You can sort on multiple fields for tie-breakers:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-request-sort.html
If you are first
The elasticsearch site has a short writeup about the benefits of the nested
type:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/mapping-nested-type.html
Basically, with object types, multiple instances of a field are flattened
as an unordered array. In your example, if you
We use a slightly inefficient technique that uses numerous buckets with
static ranges based on previously monitored heuristics and then combine the
buckets on the client side. The ranges change as you apply facets and drill
down into the documents further and the ranges change.
Cheers,
Ivan
On
The easiest way to find out how to use the Java API equivalent of a REST
call is to simply look up the RestAction class. In this case:
You can sort on multiple fields for tie-breakers:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-request-sort.html
If you are first sorting on a field, and not the score, you will need to
enable track_scores:
I am assuming that you packaged the two Java files into a jar file and
deploy it to the $ES_PATH lib directory. Elasticsearch will pick up any jar
files in that directory.
Similarities are either set to be the default for all the indices via the
index.similarity.default.type configuration
Whichever plugin you are using for the ComboAnalyzer is using the wrong
class name.
Caused by: java.lang.NoClassDefFoundError:
org/elasticsearch/ElasticSearchIllegalArgumentException
at
org.elasticsearch.index.analysis.ComboAnalyzerProvider.get(ComboAnalyzerProvider.java:50)
This class
Judging by the commits, the functionality was only deprecated, not removed.
https://github.com/elasticsearch/elasticsearch/issues/4664
I believe there are many use cases where it makes sense to boost a document
at index time. The process only occurs once instead of every time during
queries.
Correction: meant to say use *term* queries on non-analyzed fields, not
*text* queries.
On Thu, Jan 16, 2014 at 6:22 PM, Ivan Brusic i...@brusic.com wrote:
Correct. A term query does not analyze the terms, while a match query
does. Generally, you should use text queries on non-analyzed fields
I was focused on how to use a custom similarity and not how to actually
tackle the problem. In this respect, I agree with Jörg. Also, I believe the
today's release (1.0.0.RC1) has additional scripting options:
Just to clarify, when you say there's 2 fields in the row, do you really
mean that the field has two values (an array)? If so, the facet results
should look like:
alexander 1
great 1
viking 1
vinny 1
There is no easy way to achieve what you want. One way would be to great
combined keys during
You would need to combine your two filters with a bool filter:
http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-bool-filter.html
The constant score query should be at the top level and the exists/term
filters should be combined with a bool filter underneath it. The
Don't forget gateway.expected_nodes :)
Wouldn't the value increase as you add more nodes?
It will, which is precisely why the value is not computed automatically.
The value can decrease/increase over time, but the cluster does not know if
this is because it is on purpose or because of failures.
Warmer queries will affect indexing time since each new segment needs to be
warmed up. If you are going to do a large bulk load of data, it would be
advisable to disable warming.
With elasticsearch 0.90.6, it should not be necessary to register different
queries for the warmer API:
I have been accumulating a few documentation inconsistencies over the past
several months and I wanted to finally fix them and contribute them back.
Last time I made any changes was in the old markdown repository. I am
referring to the documentation under the main project (
Sounds like what you are looking for is field collapsing which is not yet
supported in elasticsearch. ETA is post 1.0 release. Perhaps there is a way
with the new aggregations framework, but I have yet to try it out.
Cheers,
Ivan
On Fri, Jan 10, 2014 at 3:13 PM, Matthew Boynes
It seems I have misunderstood the intent of the new documentation repo and
it is indeed what I am looking for. I assumed it was the WIP elasticsearch
book. Asciidoc is fairly simple, not sure why this level of complexity is
needed.
Cheers,
Ivan
On Sat, Jan 11, 2014 at 12:01 PM, Ivan Brusic i
A few things.
- I am surprised the response did not return an error. Facets work on the
document set returned by the query, so it is incorrect to add a query
section to the facet. Try it again without the query.
- Did you really use a JSON with a body section?
- The fields parameter does not
cluster.routing.allocation.disable_allocation to
reduce the need of waiting for things to rebalance.
Regards,
Mark Walkom
Infrastructure Engineer
Campaign Monitor
email: ma...@campaignmonitor.com
web: www.campaignmonitor.com
On 8 January 2014 04:41, Ivan Brusic iv...@brusic.com wrote:
Almost
That is definitely not the behavior I have ever seen with elasticsearch. If
you restart a node with allocation disabled, the restarted node will have
no shards and the shards that it should contain are marked as unassigned. I
have never seen a node reinitialize the shards it has.
Cheers,
Ivan
The CustomRestAction code you posted contains *exactly* the Java code you
need to execute the same action as the REST action.
If you want to still want to use the REST URL, you cannot use the
elasticsearch libraries. /_mastering/nodes is not a valid search type.
The action does not even execute a
BTW, I was very wrong when I mentioned that elasticsearch uses consistent
hashing. It uses modulo-based hashing, which is why the number of shards
cannot change since the modulo is fixed. Working on too many things at once
while replying. :)
On Wed, Jan 8, 2014 at 1:10 PM, InquiringMind
Here are some related links, including a video of a talk:
http://www.meetup.com/Elasticsearch-San-Francisco/events/141698772/
--
Ivan
On Tue, Jan 7, 2014 at 1:43 AM, Ümit Seren uemit.se...@gmail.com wrote:
Interesting read about elasticsearch in HipChat
://www.elasticsearch.org/guide/en/elasticsearch/reference/current/modules-advanced-scripting.html
Cheers,
Britta
On Mon, Jan 6, 2014 at 2:13 AM, Ivan Brusic i...@brusic.com wrote:
You could provide your own Similarity class as a plugin. Don't have any
sample code in front of me, but it would
Elasticsearch uses consistent hashing, so you cannot change the number of
shards for an index.
If you can reindex data, then you can create a new index with a different
number of shards and simply reindex. If your data is temporal in nature,
you can create a new index per day/week/month and these
An increase of shards will not cause an increase in sockets used. Each node
shard action is responsible for gather the responses from each shard at the
file-level before sending the response back to the client.
Since each shard is actually its own Lucene index, an increase of shards
will increase
The JSON file is used by the curl command, so in your example it should be
in the same directory in which you executed the command (current directory).
--
Ivan
On Tue, Jan 7, 2014 at 6:00 PM, ZenMaster80 sabdall...@gmail.com wrote:
Hi,
I am just starting with ElasticSearch, I would like to
I have never used non-data nodes, but in general they should primarily be
CPU-bound since it is their responsibility to gather the various shard
responses from different nodes and return a unified response to the client.
It all comes down to the amount of concurrent requests and the size of your
Analyzers are associated with fields, so in your mapping you can specify
which analyzer to use. When you query a field, elasticsearch will know
which analyzer to use (although it can be overridden). For example:
title: {type: string, analyzer: polish}
If you are using the plugin, there is no
Not yet supported:
https://github.com/elasticsearch/elasticsearch/issues/3022
Cheers,
Ivan
On Thu, Jan 2, 2014 at 4:27 AM, paul avinashpau...@gmail.com wrote:
My DATA
---
{
rankingList:[
{
value:9,
key:Academic
},
{
Judging by the one sample document, the keepalive field is not there. You
can use the missing filter to see if any documents do have that field. For
example:
curl -XPOST localhost:9200/2014010119/_count/ -d '
{
filtered: {
query: {
match_all: {}
},
filter: {
is loaded with data?
Excuse my stupid questions, but I thought the field would be created if an
index was created that had a mapping for keepalive.
Thanks
Nick
On Thursday, 2 January 2014 20:45:46 UTC, Ivan Brusic wrote:
Judging by the one sample document, the keepalive field is not there. You
AFAIK, you cannot filter on parent fields when faceting on nested documents.
Cheers,
Ivan
On Thu, Jan 2, 2014 at 2:46 PM, Nathan Moon nathannos...@gmail.com wrote:
Hi, I am using nested objects for indexing “ratings” on an object, where a
rating contains two properties: the owner and the
401 - 500 of 513 matches
Mail list logo