Without you stopping Solr itself, a solr client can remove all the
documents in an index by doing a delete-by-query with the query "*:*"
(without quotes). For XML interface clients, see
http://wiki.apache.org/solr/UpdateXmlMessage. Solrj would have another
way to do it. You'll need to do a commit after this to flush your
changes.

Alternatively, you can stop Solr and delete the whole data/ directory,
which includes the index directory. If you do this, Solr will create a
new fresh one the next time it starts up.

For backups it might be a better habit to backup the data/ directory,
rather than just the data/index directory. Assuming your schema.xml
hasn't changed, then you should be able to restore one data/ directory
with another. If you're changing your schema file, though, you need to
make sure you restore a version of that file that is consistent with
the one that you indexed with.

On Sat, Jan 24, 2009 at 5:43 PM, Johnny X <jonathanwel...@gmail.com> wrote:
>
> Thanks for the reply.
>
> I ended up fixing it by re-installing Tomcat and starting over. Searches now
> appear to work.
>
> Because I'm testing atm however, is it possible to delete the index and
> start afresh in future.
>
> At the moment I backed up the original index folder...if I just replace that
> with the current one including an index will that work...or will other parts
> of Solr recognise it's changed and as a result not work?
>
> What's the best solution for removing the index?
>
>
> Cheers.
>
>
>
> ryguasu wrote:
>>
>> I should clarify that I misspoke before; I thought you had
>> indexed="true" on Message-Id and Date, whereas you had it on
>> Message-Id and Content. It sounds like you figured this out and
>> interpreted my reply in a useful way nonetheless, though. So that's
>> good.
>>
>> The post tool should be a valid way to commit.
>>
>> As for your technique of updating the field types and reindexing the
>> documents, I think it should be fine provided you kept the field type
>> for the Message-Id field as string. If you changed it to text along
>> with the other field types, then there's a chance your "update"
>> technique might instead of the effect of inserting a duplicate copy of
>> each document, so there are two copies of each document, one
>> searchable, and one not searchable. (I'm not totally sure about this,
>> but it's a worry I would have.) That doesn't sound like what's
>> happened to you, though.
>>
>> Could the problem be that you're not specifying which field to query?
>> If you're using the standard query analyzer and the stock schema.xml,
>> then the default field name is "text", whereas you don't have a field
>> called "text" in your schema. In that setup if you want to search on
>> the Content field you need to say so explicitly, like so:
>>
>>     Content:phillip
>>
>> On Sat, Jan 24, 2009 at 7:25 AM, Johnny X <jonathanwel...@gmail.com>
>> wrote:
>>>
>>> If it helps, everything appears when I use Luke to search through the
>>> index...but the search in that returns nothing either.
>>>
>>> When I search using the admin page for the word 'Phillip' (which appears
>>> the
>>> most in all of the documents) I get the following:
>>>
>>>  <?xml version="1.0" encoding="UTF-8" ?>
>>> - <response>
>>> - <lst name="responseHeader">
>>>  <int name="status">0</int>
>>>  <int name="QTime">0</int>
>>> - <lst name="params">
>>>  <str name="indent">on</str>
>>>  <str name="start">0</str>
>>>  <str name="q">phillip</str>
>>>  <str name="rows">10</str>
>>>  <str name="version">2.2</str>
>>>  </lst>
>>>  </lst>
>>>  <result name="response" numFound="0" start="0" />
>>>  </response>
>>>
>>>
>>> Duh...?
>>>
>>>
>>>
>>> Johnny X wrote:
>>>>
>>>> They all appear in the stats admin page under the NumDocs & maxDocs
>>>> fields.
>>>>
>>>> I don't explicitly send a commit command, but my posting ends like this
>>>> (suggesting they are commited):
>>>>
>>>> SimplePostTool: POSTing file 21166.xml
>>>> SimplePostTool: POSTing file 21169.xml
>>>> SimplePostTool: COMMITting Solr index changes..
>>>>
>>>> I just tried re-posting all the documents set as "text" -- will that
>>>> update the current documents indexed? (bearing in mind the unique key,
>>>> message-id, will be included again)
>>>>
>>>> When I try searching I still get 0 results for anything included in the
>>>> message-id and content fields, both of which should be indexed and
>>>> returning results...
>>>>
>>>>
>>>> Cheers for any help!
>>>>
>>>>
>>>> ryguasu wrote:
>>>>>
>>>>> These might be obvious, but:
>>>>>
>>>>> * I assume you did a Solr commit command after indexing, right?
>>>>>
>>>>> * If you are using the fieldtype definitions from the default
>>>>> schema.xml, then your "string" fields are not being analyzed, which
>>>>> means you should expect search results only if you enter the entire,
>>>>> exact value of one of the Message-ID or Date fields in your query. Is
>>>>> that your intention?
>>>>>
>>>>> And yes, your analysis of "stored" seems correct. Stored fields are
>>>>> those whose values you need back at query time, and indexed fields are
>>>>> those you can do queries on. For a few complications, see
>>>>> http://wiki.apache.org/solr/FieldOptionsByUseCase
>>>>>
>>>>> On Fri, Jan 23, 2009 at 8:04 PM, Johnny X <jonathanwel...@gmail.com>
>>>>> wrote:
>>>>>>
>>>>>> I've indexed my XML using the below in the schema:
>>>>>>
>>>>>>   <field name="Message-ID" type="string" indexed="true" stored="true"
>>>>>> required="true"/>
>>>>>>   <field name="Date" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="From" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="To" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="Subject" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="Mime-Version" type="string" indexed="false"
>>>>>> stored="true"/>
>>>>>>   <field name="Content-Type" type="string" indexed="false"
>>>>>> stored="true"/>
>>>>>>   <field name="Content-Transfer-Encoding" type="string"
>>>>>> indexed="false"
>>>>>> stored="true"/>
>>>>>>   <field name="X-From" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="X-To" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="X-cc" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="X-bcc" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="X-Folder" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="X-Origin" type="string" indexed="false" stored="true"/>
>>>>>>   <field name="X-FileName" type="string" indexed="false"
>>>>>> stored="true"/>
>>>>>>   <field name="Content" type="string" indexed="true" stored="true"/>
>>>>>>
>>>>>>  <uniqueKey>Message-ID</uniqueKey>
>>>>>>
>>>>>> However searching via the Message-ID or Content fields returns 0.
>>>>>> Using
>>>>>> Luke
>>>>>> I can still see these fields are stored however.
>>>>>>
>>>>>> Out of interest, by setting the other fields to just "stored=true",
>>>>>> can
>>>>>> they
>>>>>> be returned in a query as part of a search?
>>>>>>
>>>>>>
>>>>>> Cheers.
>>>>>> --
>>>>>> View this message in context:
>>>>>> http://www.nabble.com/Results-not-appearing-tp21637069p21637069.html
>>>>>> Sent from the Solr - User mailing list archive at Nabble.com.
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>
>>> --
>>> View this message in context:
>>> http://www.nabble.com/Results-not-appearing-tp21637069p21641692.html
>>> Sent from the Solr - User mailing list archive at Nabble.com.
>>>
>>>
>>
>>
>
> --
> View this message in context: 
> http://www.nabble.com/Results-not-appearing-tp21637069p21647649.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>
>

Reply via email to