directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefan Seelmann <seelm...@apache.org>
Subject Re: HBase partition integration in trunks ?
Date Tue, 16 Aug 2011 13:20:29 GMT
On Tue, Aug 16, 2011 at 3:14 PM, Kiran Ayyagari <kayyagari@apache.org> wrote:

>> If you think about what is done when we add an entry in the current code
>> base :
>> - add the entry in the MasterTable
>> - add the RDN/parent into the RdnIndex
>> - update the one-level index with the newly added entry reference,
>> increasing the number of children for the parent
>> - for each RDN in the parent, update the sub-level index with the newly
>> added entry reference, increasing the number of children for the parent
>>
>> If we compare what we would do if we remove the one-level/sub-level index :
>> - add the entry in the MasterTable
>> - add the RDN/parent into the RdnIndex
>> - for each RDN in the parent, update the rdn index with the newly added
>> entry reference, increasing the number of children for the parent
>>
> this count needs to be updated in more than one ParentAndRdn entry(for
> sublevel indexing feature only) depending on the level
> at which the child entry was added(cause each entry in the ancestor
> chain is a valid base to perform sublevel search)

Correct.

> and also when it comes to fetch all the sublevel entries we need to
> scan recursively.

Right. I used depth-first tree traversal. So in that case the maximum
number of open cursors is the max. depth of the DIT.

Kind Regards,
Stefan

Mime
View raw message