changing int to long - does it definitely require a reindex?

[Ty]

I messed up.  In my 50+ million document index, I have a field in my
schema.xml that is of type "int".  I should have made it a long; documents
that have a field that overflows that integer aren't being indexed.

Does changing this field type absolutely, positively require a re-index?
 What would happen if I simply changed the data type without re-indexing
old documents?

Thanks,
Ty

using HTTP caching with shards in Solr 4.3

[Ty]

Hello,

I'd like to take advantage of Solr's HTTP caching feature (httpCaching
never304="false" in solrconfig.xml)..  It is behaving as expected when I do
a standard query against a Solr instance and then repeat it: I receive an
HTTP304 (not modified) response.

However, when using the "shards" functionality, I seem to be unable to get
the HTTP304 functionality.  When sending a request to a Solr instance that
includes other Solr instances in the "shards" parameter, a GET request is
sent to the original Solr instance, but it turns around and sends POST
requests to the Solr instances referenced in "shards".  Since POST requests
cannot generate a 304, I seem to be unable to use HTTP caching with shards.

Is there a way to make the original Solr instance query the shards with a
GET method?  Or some other way I can leverage HTTP caching when using
shards?

Thanks,
Ty

Dismax and Grouping query

[Ty Hahn]

Hi,

I've tried to use grouping query on DisMaxRequestHandler without success.
When I sent grouping query in Solr Admin, I could see parens of query
escaped in 'querystring' line with debugQuery On.
Is this the cause of the failure?

e.g.
When I send query like +(lucene solr),
I can see following line in the result page.
+\(lucene solr\)

When I tried this with StandardRequestHandler, parens of the query were not
escaped. And the query was successfully answered.
Digging into the source of Solr, I could find the following line at
DisMaxRequestHandler.java.
userQuery = U.partialEscape(U.stripUnbalancedQuotes(userQuery)).toString();
And partialEscape function seems to carry out the escaping.

So... Can I carry out grouping query on DisMaxRequestHandler?
If so, should I use special character for grouping in stead of parens?

I'm pretty new on Solr. Any reply will help.
Thanks in advance.

Re: Dismax and Grouping query

[Ty Hahn]

Thanks, Hoss.
It seems that I should think other options.
Thanks again.


On 9/29/07, Chris Hostetter <[EMAIL PROTECTED]> wrote:
>
>
> : I've tried to use grouping query on DisMaxRequestHandler without
> success.
>
> : e.g.
> : When I send query like +(lucene solr),
> : I can see following line in the result page.
> : +\(lucene solr\)
>
> the dismax handler does not consider parens to be special characters.  if
> it did, it's not clear what the semantics would be of a query like...
> q=A +(B C)qf=X Y Z
> ..when building the query structure ... what happens if X:B exists and Y:C
> exists? is that considered a match?
>
> Generally, the "mm" param is used to indicate how many of the query terms
> (that don't have a + or - prefix) are required, or you can explicitly
> require/prohibit a term using + or -, but there is no way to require that
> one of N sub terms is required (prohibiting any of N sub terms is easy,
> just prohibit them all individually)
>
>
>
> -Hoss
>
>