efficient sort by title (multi word field)

[Gili Nachum <gi...@gmail.com>]

Hi, wanted to make sure I'm implementing sort in an efficient way...

I need to allow users to sort by documents' title field. A title can
contain 1-20 words.
Title examples: "new project meeting minutes - Oct 2015 - new chance on the
horizon" or "how to create a wonderful presentation".

I'm already indexing title as a TextField, and I not comfortable with
indexing it again as an extra StrField field I'll need plus the extra
FieldCache memory.
I can probably avoid the FieldCache by using docValues to reduce mem usage.

*But... is there a more efficient way to provide this sort? perhaps
something that takes advantage of title field is a chain of words with
whitespaces between words?*

My index is 100's of millions of documents over 8 shards.

Thanks.

Re: efficient sort by title (multi word field)

[Alessandro Benedetti <be...@gmail.com>]

Hi Gili,
if you want to use sorting you will need to index the extra string field.
And I suggest you to set only DocValues on that field.
Doc values are not compatible with textual analysed fields and neither sort
is.

[1]
https://cwiki.apache.org/confluence/display/solr/Common+Query+Parameters#CommonQueryParameters-ThesortParameter

Cheers

On 6 October 2015 at 19:27, Gili Nachum <gi...@gmail.com> wrote:

> Hi, wanted to make sure I'm implementing sort in an efficient way...
>
> I need to allow users to sort by documents' title field. A title can
> contain 1-20 words.
> Title examples: "new project meeting minutes - Oct 2015 - new chance on the
> horizon" or "how to create a wonderful presentation".
>
> I'm already indexing title as a TextField, and I not comfortable with
> indexing it again as an extra StrField field I'll need plus the extra
> FieldCache memory.
> I can probably avoid the FieldCache by using docValues to reduce mem usage.
>
> *But... is there a more efficient way to provide this sort? perhaps
> something that takes advantage of title field is a chain of words with
> whitespaces between words?*
>
> My index is 100's of millions of documents over 8 shards.
>
> Thanks.
>



-- 
--------------------------

Benedetti Alessandro
Visiting card - http://about.me/alessandro_benedetti
Blog - http://alexbenedetti.blogspot.co.uk

"Tyger, tyger burning bright
In the forests of the night,
What immortal hand or eye
Could frame thy fearful symmetry?"

William Blake - Songs of Experience -1794 England