Find problem in eXtremeDB

标签：

class table1

{

char<8>    f1;

char<80>  f2;

uint4        f3;

uint4        f4;

double     f5;

uint4        f6;

uniquetree<f1,f2> Idx1;

tree<f6>Idx2; 

};

Steps：

step1：find（Idx1=（a，b）），find the records;

step2：put（f6,c）;

step3：find（Idx1=（a，b）），did not find the record instep 1.

The record should be there, what happened when the putmethod is used, the index is changed after _put method?

If a class has one or more indexes, thenthe field(s) on which the index is defined

will have an index component (hash tableentry or tree node) in addition to the

actual field value. The index component isnot updated when the field’s _put()

function is called, but rather when theREAD_WRITE transaction containing this

update is committed. Or, alternatively, a_checkpoint() function can be called

to explicitly update the index componentsfor this object. The _checkpoint()

function completes the object’s update before the transaction is committed,

however if the application decides to rollback the current transaction, all the

updates for the object including indexcomponents are discarded. (Committing a

transaction implicitly checkpoints all theobjects created, updated or deleted in the

transaction.)

Updating f6causes the object to be removed from all indexes.  Using _checkpoint()after _put() is the correct approach.

Find problem in eXtremeDB

标签：

原文地址：http://blog.csdn.net/longlongago7777/article/details/44492995