Unique dictionary field question

[march]

We detected an issue due to the fact that a field (correlation_id on the cmdb_ci table) has the dictionary attribute "unique" set to "true". Comparing with a brand new system, we discovered that, OoB, this attribute is set to "false".

We therefore decided to go OoB and set it to false on our instance.

But now, despite this "unique" attribute is set to false, when we test it and try to enter in this field a duplicate value, we still get the following message:

java.sql.BatchUpdateException: Duplicate entry '42 31 54 a5 49 98 59 1f-2c 65 96 74 70 c6 c3 c3' for key 'correlation_id'

What else needs to be done to have this change in effect?

Thanks

Marcus

[Omkar Mone]

Hi 

Try clearing the cache and perform the steps again.

Please let me know if it worked.

Thanks.

[march]

Thanks for your answer,

I forgot to mention, but i did this already (cache.do) and this not changed anything.

[Mark Stanger]

You might try checking the 'Unique' attribute on another field (like sys_id) that should be unique anyway.  It's kind of a shot in the dark, but it might send the system whatever message it needs to re-evaluate things for that table.  Using sys_id for this test shouldn't hurt anything because it should be unique anyway.  You could also create a custom field to test and then delete it when you're done.

[march]

Thanks Mark, i just created a new field as you propose but it did not worked.

In fact i flagged that new field as unique, that makes it indeed unique.

I then unflag it and have the same issue as for the correlation_id.

My understanding is that checking that flag will create a "UNIQUE" DB index. So, is there a way to drop index?