Re: Deleting document on wrong shard?

[Thomas Corthals]

Hi Walter

Deleting multiple IDs at once with JSON is mentioned here:
https://solr.apache.org/guide/solr/latest/indexing-guide/indexing-with-update-handlers.html#sending-json-update-commands

Or a list of document IDs:
>
> { "delete":["id1","id2"] }
>
>
Thomas

Op wo 24 mei 2023 om 22:23 schreef Walter Underwood :

> It works! Thanks so much.
>
> I’m using XML update format because the JSON format for sending multiple
> IDs for deletion is not documented anywhere I could find. It was easier to
> just generate XML instead of continuing to search for documentation.
>
> This does the trick:
>
>  route="shard25_2">datalake_FPD_163298_3RGR-V090-006H-40F0-0-00
>
> wunder
> Walter Underwood
> wun...@wunderwood.org
> http://observer.wunderwood.org/  (my blog)
>

join query parser performance

[Ron Haines]

I've been using the 'join' query parser to 'filter out' related documents
that should not be part of the result set.  Functionally, it is working
fine.  However, when we throw a 'real' level of customer traffic at it, it
pretty much brings Solr to its knees.  CPU increases ALOT.  Close to 3X,
when I enable this feature in our system.  Solr response times shoot up,
and thread counts shoot up.  Before I 'give up' on the join query parser, I
thought I'd seek some advice here.

So, when this feature is enabled, this negative &fq gets added:
-{!join fromIndex=primary_rollup from=group_id_mv to=group_member_id
score=none}${q}

The 'local' collection size is about 27 million docs, but the number of
docs that actually contain a 'group_member_id' is only about 125k.  And, in
the 'fromIndex' collection, there are only 80k documents in that
collection, and they all have the 'group_id_mv' field.  The 'fromIndex'
collection is a single shard, with a replica on each shard of the local
collection.  The local collection only has about 300k docs per shard, at 96
shards.

I guess I'm just trying to understand why this appears to be causing such
problems for Solr, as the amount of work (the # of documents involved)
seems relatively small.

I hope I'm missing something...
Thanks for any input.

Re: join query parser performance

[Joel Bernstein]

If you are using a recent version of Solr try adding the parameter

method=topLevelDV

Let us know how this effects performance in your use case.

What matters most here is the number of documents the from side of join
matches.





Joel Bernstein
http://joelsolr.blogspot.com/


On Thu, May 25, 2023 at 8:52 AM Ron Haines  wrote:

> I've been using the 'join' query parser to 'filter out' related documents
> that should not be part of the result set.  Functionally, it is working
> fine.  However, when we throw a 'real' level of customer traffic at it, it
> pretty much brings Solr to its knees.  CPU increases ALOT.  Close to 3X,
> when I enable this feature in our system.  Solr response times shoot up,
> and thread counts shoot up.  Before I 'give up' on the join query parser, I
> thought I'd seek some advice here.
>
> So, when this feature is enabled, this negative &fq gets added:
> -{!join fromIndex=primary_rollup from=group_id_mv to=group_member_id
> score=none}${q}
>
> The 'local' collection size is about 27 million docs, but the number of
> docs that actually contain a 'group_member_id' is only about 125k.  And, in
> the 'fromIndex' collection, there are only 80k documents in that
> collection, and they all have the 'group_id_mv' field.  The 'fromIndex'
> collection is a single shard, with a replica on each shard of the local
> collection.  The local collection only has about 300k docs per shard, at 96
> shards.
>
> I guess I'm just trying to understand why this appears to be causing such
> problems for Solr, as the amount of work (the # of documents involved)
> seems relatively small.
>
> I hope I'm missing something...
> Thanks for any input.
>

Re: join query parser performance

[Joel Bernstein]

One thing to understand about the topLevelDv approach is you'll need to
warm both sides of the join. You can do this by adding a  static warming
query that facets on 'group_id_mv'  and 'group_member_id' in both
collections.


Joel Bernstein
http://joelsolr.blogspot.com/


On Thu, May 25, 2023 at 9:17 AM Joel Bernstein  wrote:

> If you are using a recent version of Solr try adding the parameter
>
> method=topLevelDV
>
> Let us know how this effects performance in your use case.
>
> What matters most here is the number of documents the from side of join
> matches.
>
>
>
>
>
> Joel Bernstein
> http://joelsolr.blogspot.com/
>
>
> On Thu, May 25, 2023 at 8:52 AM Ron Haines  wrote:
>
>> I've been using the 'join' query parser to 'filter out' related documents
>> that should not be part of the result set.  Functionally, it is working
>> fine.  However, when we throw a 'real' level of customer traffic at it, it
>> pretty much brings Solr to its knees.  CPU increases ALOT.  Close to 3X,
>> when I enable this feature in our system.  Solr response times shoot up,
>> and thread counts shoot up.  Before I 'give up' on the join query parser,
>> I
>> thought I'd seek some advice here.
>>
>> So, when this feature is enabled, this negative &fq gets added:
>> -{!join fromIndex=primary_rollup from=group_id_mv to=group_member_id
>> score=none}${q}
>>
>> The 'local' collection size is about 27 million docs, but the number of
>> docs that actually contain a 'group_member_id' is only about 125k.  And,
>> in
>> the 'fromIndex' collection, there are only 80k documents in that
>> collection, and they all have the 'group_id_mv' field.  The 'fromIndex'
>> collection is a single shard, with a replica on each shard of the local
>> collection.  The local collection only has about 300k docs per shard, at
>> 96
>> shards.
>>
>> I guess I'm just trying to understand why this appears to be causing such
>> problems for Solr, as the amount of work (the # of documents involved)
>> seems relatively small.
>>
>> I hope I'm missing something...
>> Thanks for any input.
>>
>

Re: join query parser performance

[Deepak Goel]

Ron,

please post actual figures:

1. Cpu, mem, disk, network utilisation
2. Response times
3. Load
4. Hardware config of server
5. Software config of server


On Thu, 25 May 2023, 18:51 Joel Bernstein,  wrote:

> One thing to understand about the topLevelDv approach is you'll need to
> warm both sides of the join. You can do this by adding a  static warming
> query that facets on 'group_id_mv'  and 'group_member_id' in both
> collections.
>
>
> Joel Bernstein
> http://joelsolr.blogspot.com/
>
>
> On Thu, May 25, 2023 at 9:17 AM Joel Bernstein  wrote:
>
> > If you are using a recent version of Solr try adding the parameter
> >
> > method=topLevelDV
> >
> > Let us know how this effects performance in your use case.
> >
> > What matters most here is the number of documents the from side of join
> > matches.
> >
> >
> >
> >
> >
> > Joel Bernstein
> > http://joelsolr.blogspot.com/
> >
> >
> > On Thu, May 25, 2023 at 8:52 AM Ron Haines  wrote:
> >
> >> I've been using the 'join' query parser to 'filter out' related
> documents
> >> that should not be part of the result set.  Functionally, it is working
> >> fine.  However, when we throw a 'real' level of customer traffic at it,
> it
> >> pretty much brings Solr to its knees.  CPU increases ALOT.  Close to 3X,
> >> when I enable this feature in our system.  Solr response times shoot up,
> >> and thread counts shoot up.  Before I 'give up' on the join query
> parser,
> >> I
> >> thought I'd seek some advice here.
> >>
> >> So, when this feature is enabled, this negative &fq gets added:
> >> -{!join fromIndex=primary_rollup from=group_id_mv to=group_member_id
> >> score=none}${q}
> >>
> >> The 'local' collection size is about 27 million docs, but the number of
> >> docs that actually contain a 'group_member_id' is only about 125k.  And,
> >> in
> >> the 'fromIndex' collection, there are only 80k documents in that
> >> collection, and they all have the 'group_id_mv' field.  The 'fromIndex'
> >> collection is a single shard, with a replica on each shard of the local
> >> collection.  The local collection only has about 300k docs per shard, at
> >> 96
> >> shards.
> >>
> >> I guess I'm just trying to understand why this appears to be causing
> such
> >> problems for Solr, as the amount of work (the # of documents involved)
> >> seems relatively small.
> >>
> >> I hope I'm missing something...
> >> Thanks for any input.
> >>
> >
>

Redindex Collection API stalled

[Rajani Maski]

Hi Solr Users,

   The reindex collection api stops importing after it imports around 20%
of total docs from source collection, how to restart it? I could find
"start", "abort" and "status" commands on the documentation page, but there
is no command to restart from where it left. The status is running however
it is stalled. Following is the api I used to start it.  Any suggestions on
how to restart from where it left, maybe update the checkpoint collection?



http://host:8983/solr/admin/collections?action=REINDEXCOLLECTION&name=v9&numShards=12&collection.configName=v9-schema&replicationFactor=1

Re: join query parser performance

[Andy Lester]

> On May 25, 2023, at 7:51 AM, Ron Haines  wrote:
> 
> So, when this feature is enabled, this negative &fq gets added:
> -{!join fromIndex=primary_rollup from=group_id_mv to=group_member_id
> score=none}${q}


Can we see collection definitions of both the source collection and the join? 
Also, a sample query, not just the one parameter? Also, how often are either of 
these collections updated? One thing that killed off an entire project that we 
were doing was that the join table was getting updated about once a minute, and 
this destroyed all our caching, and made the queries we wanted to do unusable.


Thanks,
Andy

Re: Redindex Collection API stalled

[Ed K]

 it has happened to me in the past and I've had to restart it from beginning.

On Thursday, May 25, 2023 at 06:46:04 AM PDT, Rajani Maski 
 wrote:  
 
 Hi Solr Users,

  The reindex collection api stops importing after it imports around 20%
of total docs from source collection, how to restart it? I could find
"start", "abort" and "status" commands on the documentation page, but there
is no command to restart from where it left. The status is running however
it is stalled. Following is the api I used to start it.  Any suggestions on
how to restart from where it left, maybe update the checkpoint collection?



http://host:8983/solr/admin/collections?action=REINDEXCOLLECTION&name=v9&numShards=12&collection.configName=v9-schema&replicationFactor=1
  

Re: Deleting document on wrong shard?

[Walter Underwood]

Hmm. That doesn’t seem consistent with the format change added in SOLR-5980, 
where each ID can have a _route_ attribute.

{ "id":"ID" , "_route_":"route”}

https://issues.apache.org/jira/browse/SOLR-5890?focusedCommentId=14234208&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-14234208

wunder
Walter Underwood
wun...@wunderwood.org
http://observer.wunderwood.org/  (my blog)

> On May 25, 2023, at 12:19 AM, Thomas Corthals  wrote:
> 
> Hi Walter
> 
> Deleting multiple IDs at once with JSON is mentioned here:
> https://solr.apache.org/guide/solr/latest/indexing-guide/indexing-with-update-handlers.html#sending-json-update-commands
> 
> Or a list of document IDs:
>> 
>> { "delete":["id1","id2"] }
>> 
>> 
> Thomas
> 
> Op wo 24 mei 2023 om 22:23 schreef Walter Underwood :
> 
>> It works! Thanks so much.
>> 
>> I’m using XML update format because the JSON format for sending multiple
>> IDs for deletion is not documented anywhere I could find. It was easier to
>> just generate XML instead of continuing to search for documentation.
>> 
>> This does the trick:
>> 
>> > route="shard25_2">datalake_FPD_163298_3RGR-V090-006H-40F0-0-00
>> 
>> wunder
>> Walter Underwood
>> wun...@wunderwood.org
>> http://observer.wunderwood.org/  (my blog)
>> 

Re: Redindex Collection API stalled

[ufuk yılmaz]

Are you specifying a list of field names when calling the api?

-ufuk yilmaz

—

> On 25 May 2023, at 17:47, Ed K  wrote:
> 
> it has happened to me in the past and I've had to restart it from beginning.
> 
>On Thursday, May 25, 2023 at 06:46:04 AM PDT, Rajani Maski 
>  wrote:  
> 
> Hi Solr Users,
> 
>   The reindex collection api stops importing after it imports around 20%
> of total docs from source collection, how to restart it? I could find
> "start", "abort" and "status" commands on the documentation page, but there
> is no command to restart from where it left. The status is running however
> it is stalled. Following is the api I used to start it.  Any suggestions on
> how to restart from where it left, maybe update the checkpoint collection?
> 
> 
> 
> http://host:8983/solr/admin/collections?action=REINDEXCOLLECTION&name=v9&numShards=12&collection.configName=v9-schema&replicationFactor=1

Question on Solr filter syntax

[Wei]

  Send on behalf of my colleague as his email seems not showing up in the
list.

Hi Team,


I was reading about special filter syntax in Solr documentation -

https://solr.apache.org/guide/8_3/the-standard-query-parser.html#differences-between-lucenes-classic-query-parser-and-solrs-standard-query-parser


Support for a special filter(…) syntax to indicate that some query
clauses should be cached in the filter cache (as a constant score
boolean query). This allows sub-queries to be cached and re-used in
other queries. For example inStock:true will be cached and re-used in
all three of the queries below:

q=features:songs OR filter(inStock:true)


My question is how do I disable cache for part of OR filter query and
enable for part of OR filter query?


For e.g. - Current FQ -

fq: planName:all_features OR zipCode:1234 (in this case entire FQ is cached)


I don't want to cache the entire FQ given it's very unique , but I
want to cache the first part (plan_name:all_features) to improve
performance . How do I achieve this?


My understanding is the following should work? Is this correct?

fq: {!cache=false}(filter(planName:all_features) OR zipCode:1234)


Will the above syntax result in planName:all_features to be cached in
Filter Cache? Otherwise how can I achieve this?


I am on Solr 8.4


Thanks,

Yash

Question on solr filter syntax

[Wei]

Send on behalf of my colleague as his email seems not showing up in the
list.

Hi Team,


I was reading about special filter syntax in Solr documentation -

https://solr.apache.org/guide/8_3/the-standard-query-parser.html#differences-between-lucenes-classic-query-parser-and-solrs-standard-query-parser


Support for a special filter(…) syntax to indicate that some query
clauses should be cached in the filter cache (as a constant score
boolean query). This allows sub-queries to be cached and re-used in
other queries. For example inStock:true will be cached and re-used in
all three of the queries below:

q=features:songs OR filter(inStock:true)


My question is how do I disable cache for part of OR filter query and
enable for part of OR filter query?


For e.g. - Current FQ -

fq: planName:all_features OR zipCode:1234 (in this case entire FQ is cached)


I don't want to cache the entire FQ given it's very unique , but I
want to cache the first part (plan_name:all_features) to improve
performance . How do I achieve this?


My understanding is the following should work? Is this correct?

fq: {!cache=false}(filter(planName:all_features) OR zipCode:1234)


Will the above syntax result in planName:all_features to be cached in
Filter Cache? Otherwise how can I achieve this?


I am on Solr 8.4


Thanks,

Yash

Re: Redindex Collection API stalled

[Rajani Maski]

no, whatever the default is, assuming it is fl=*  Would it help if it lists
field names? It will be a large list, about 50-60 fields so leaving it as
default. All the fields are stored.

On Thu, May 25, 2023 at 11:09 AM ufuk yılmaz 
wrote:

> Are you specifying a list of field names when calling the api?
>
> -ufuk yilmaz
>
> —
>
> > On 25 May 2023, at 17:47, Ed K  wrote:
> >
> > it has happened to me in the past and I've had to restart it from
> beginning.
> >
> >On Thursday, May 25, 2023 at 06:46:04 AM PDT, Rajani Maski <
> rajinima...@gmail.com> wrote:
> >
> > Hi Solr Users,
> >
> >   The reindex collection api stops importing after it imports around 20%
> > of total docs from source collection, how to restart it? I could find
> > "start", "abort" and "status" commands on the documentation page, but
> there
> > is no command to restart from where it left. The status is running
> however
> > it is stalled. Following is the api I used to start it.  Any suggestions
> on
> > how to restart from where it left, maybe update the checkpoint
> collection?
> >
> >
> >
> >
> http://host:8983/solr/admin/collections?action=REINDEXCOLLECTION&name=v9&numShards=12&collection.configName=v9-schema&replicationFactor=1
>
>

Vector Search Parser(!vp) : calculate cosine

[Rajani Maski]

Hi Solr Users,

   Is there a vector search parser that allows computing distance between a
document's vector field and query vector passed as query param?  The jiras
SOLR-14397   SOLR-12890
  suggests vector search
parser "!vp" and the other "cosine" as functions but I couldn't get it to
work on solr 9.1.1 version, is this supported yet, one of the jira state is
still "open"?

What are other alternatives to calculate cosine between vector fields in
solr vs query vector sent as query param?


Thanks,
Rajani

Re: Redindex Collection API stalled

[ufuk yılmaz]

Afaik reindexcollection command uses streaming expressions and export handler 
under the hood, so the fields need to have docValues. I need to check the 
source code to see if it also works with just stored fields. 

If you have copyField s in your schema, it may also interfere because they 
would be sent two times to the target field

I don’t know, reindexcollection only works for me if I exclude copyFields and 
only use fields with docValues. Also if your index is very large (100s gb) it 
also times out silently. 

-ufuk yilmaz
—

> On 25 May 2023, at 21:27, Rajani Maski  wrote:
> 
> no, whatever the default is, assuming it is fl=*  Would it help if it lists
> field names? It will be a large list, about 50-60 fields so leaving it as
> default. All the fields are stored.
> 
> On Thu, May 25, 2023 at 11:09 AM ufuk yılmaz 
> wrote:
> 
>> Are you specifying a list of field names when calling the api?
>> 
>> -ufuk yilmaz
>> 
>> —
>> 
>>> On 25 May 2023, at 17:47, Ed K  wrote:
>>> 
>>> it has happened to me in the past and I've had to restart it from
>> beginning.
>>> 
>>>   On Thursday, May 25, 2023 at 06:46:04 AM PDT, Rajani Maski <
>> rajinima...@gmail.com> wrote:
>>> 
>>> Hi Solr Users,
>>> 
>>>  The reindex collection api stops importing after it imports around 20%
>>> of total docs from source collection, how to restart it? I could find
>>> "start", "abort" and "status" commands on the documentation page, but
>> there
>>> is no command to restart from where it left. The status is running
>> however
>>> it is stalled. Following is the api I used to start it.  Any suggestions
>> on
>>> how to restart from where it left, maybe update the checkpoint
>> collection?
>>> 
>>> 
>>> 
>>> 
>> http://host:8983/solr/admin/collections?action=REINDEXCOLLECTION&name=v9&numShards=12&collection.configName=v9-schema&replicationFactor=1
>> 
>> 

Re: Redindex Collection API stalled

[rajani m]

oh ok, it may be the silent timeout then.  It has reindexed over a million
records that mostly have stored fields. It stalled after that, the source
collection has 20m records so it may be a large collection and silent
timeout.  The endpoint created a checkpoint collection that has one
document in it, thinking a manual update to it may allow it to reset to
where it can restart from? It does not allow writing to an existing
collection otherwise large collection reindex with q set query by date
could have been an alternative.



On Thu, May 25, 2023 at 2:55 PM ufuk yılmaz 
wrote:

> Afaik reindexcollection command uses streaming expressions and export
> handler under the hood, so the fields need to have docValues. I need to
> check the source code to see if it also works with just stored fields.
>
> If you have copyField s in your schema, it may also interfere because they
> would be sent two times to the target field
>
> I don’t know, reindexcollection only works for me if I exclude copyFields
> and only use fields with docValues. Also if your index is very large (100s
> gb) it also times out silently.
>
> -ufuk yilmaz
> —
>
> > On 25 May 2023, at 21:27, Rajani Maski  wrote:
> >
> > no, whatever the default is, assuming it is fl=*  Would it help if it
> lists
> > field names? It will be a large list, about 50-60 fields so leaving it as
> > default. All the fields are stored.
> >
> > On Thu, May 25, 2023 at 11:09 AM ufuk yılmaz  >
> > wrote:
> >
> >> Are you specifying a list of field names when calling the api?
> >>
> >> -ufuk yilmaz
> >>
> >> —
> >>
> >>> On 25 May 2023, at 17:47, Ed K  wrote:
> >>>
> >>> it has happened to me in the past and I've had to restart it from
> >> beginning.
> >>>
> >>>   On Thursday, May 25, 2023 at 06:46:04 AM PDT, Rajani Maski <
> >> rajinima...@gmail.com> wrote:
> >>>
> >>> Hi Solr Users,
> >>>
> >>>  The reindex collection api stops importing after it imports around 20%
> >>> of total docs from source collection, how to restart it? I could find
> >>> "start", "abort" and "status" commands on the documentation page, but
> >> there
> >>> is no command to restart from where it left. The status is running
> >> however
> >>> it is stalled. Following is the api I used to start it.  Any
> suggestions
> >> on
> >>> how to restart from where it left, maybe update the checkpoint
> >> collection?
> >>>
> >>>
> >>>
> >>>
> >>
> http://host:8983/solr/admin/collections?action=REINDEXCOLLECTION&name=v9&numShards=12&collection.configName=v9-schema&replicationFactor=1
> >>
> >>
>
>

Re: Vector Search Parser(!vp) : calculate cosine

[kumar gaurav]

HI Rajani

Read here :
https://solr.apache.org/guide/solr/latest/query-guide/dense-vector-search.html

The knn query parser can be used .

&q={!knn f=vector topK=10}[1.0, 2.0, 3.0, 4.0]


Thanks
Kumar Gaurav



On Fri, 26 May 2023 at 00:14, Rajani Maski  wrote:

> Hi Solr Users,
>
>Is there a vector search parser that allows computing distance between a
> document's vector field and query vector passed as query param?  The jiras
> SOLR-14397   SOLR-12890
>   suggests vector search
> parser "!vp" and the other "cosine" as functions but I couldn't get it to
> work on solr 9.1.1 version, is this supported yet, one of the jira state is
> still "open"?
>
> What are other alternatives to calculate cosine between vector fields in
> solr vs query vector sent as query param?
>
>
> Thanks,
> Rajani
>

Re: join query parser performance

[Mikhail Khludnev]

Ron, how often both indices are updated? Presumably if they are static,
filter cache may help.
It's worth making sure that the app gives a chance to filter cache.;
To better understand the problem it is worth taking a few treadumps under
load: a deep stack gives a clue for hotspot (or just take a sampling
profile). Once we know the hot spot we can think about a workaround.
https://issues.apache.org/jira/browse/SOLR-16717 about sharding "fromIndex"
https://issues.apache.org/jira/browse/SOLR-16242 about keeping "local/to"
index cache when fromIndex is updated.

On Thu, May 25, 2023 at 5:01 PM Andy Lester  wrote:

>
>
> > On May 25, 2023, at 7:51 AM, Ron Haines  wrote:
> >
> > So, when this feature is enabled, this negative &fq gets added:
> > -{!join fromIndex=primary_rollup from=group_id_mv to=group_member_id
> > score=none}${q}
>
>
> Can we see collection definitions of both the source collection and the
> join? Also, a sample query, not just the one parameter? Also, how often are
> either of these collections updated? One thing that killed off an entire
> project that we were doing was that the join table was getting updated
> about once a minute, and this destroyed all our caching, and made the
> queries we wanted to do unusable.
>
>
> Thanks,
> Andy



-- 
Sincerely yours
Mikhail Khludnev

Re: Vector Search Parser(!vp) : calculate cosine

[rajani m]

That also invokes knn, correct? I just need the vector math response, a
cosine value of the query vector and document vector. For example, the
streaming api endpoint vector-math.html#dot-product-and-cosine-similarity

looks like one however it is a streaming api, it would have been nice if
this was also supported by standard query parser or a function query.



On Thu, May 25, 2023 at 3:34 PM kumar gaurav  wrote:

> HI Rajani
>
> Read here :
>
> https://solr.apache.org/guide/solr/latest/query-guide/dense-vector-search.html
>
> The knn query parser can be used .
>
> &q={!knn f=vector topK=10}[1.0, 2.0, 3.0, 4.0]
>
>
> Thanks
> Kumar Gaurav
>
>
>
> On Fri, 26 May 2023 at 00:14, Rajani Maski  wrote:
>
> > Hi Solr Users,
> >
> >Is there a vector search parser that allows computing distance
> between a
> > document's vector field and query vector passed as query param?  The
> jiras
> > SOLR-14397 
> SOLR-12890
> >   suggests vector
> search
> > parser "!vp" and the other "cosine" as functions but I couldn't get it to
> > work on solr 9.1.1 version, is this supported yet, one of the jira state
> is
> > still "open"?
> >
> > What are other alternatives to calculate cosine between vector fields in
> > solr vs query vector sent as query param?
> >
> >
> > Thanks,
> > Rajani
> >
>

Re: Question on Solr filter syntax

[Shawn Heisey]

On 5/25/23 11:00, Wei wrote:

For e.g. - Current FQ -

fq: planName:all_features OR zipCode:1234 (in this case entire FQ is cached)

I don't want to cache the entire FQ given it's very unique , but I
want to cache the first part (plan_name:all_features) to improve
performance . How do I achieve this?


I am about 90 percent sure this is not possible.  In the filterCache, 
the entire text of the filter is used as the key for each entry in the 
cache.


I'm not familiar with that part of the code, but I bet that even if it 
IS possible, and it might not be, it would take a near-complete rewrite 
of Solr's filter query code.


Thanks,
Shawn