On 26/09/12 10:42, Suneetha wrote:
Nested Transactions: This required in order to roll back
transactions, in case of any problem in between.. say for example
operation 'save', where lot of sub records transactions have to be
saved before saving the main record transaction..
Observation: Nested transactions aren't necessary for that - wait until
the top level transaction commits.
SDB: SDB as we are using IBM DB2 database as a storage and SDB
supports this database, but not TDB..
One more thing just out of curiosity: What is the abbreviation of SDB
and TDB?:-)
S = SQL ... or SPARQL.
T = Tuple.
Andy
--- On Wed, 26/9/12, Andy Seaborne <[email protected]> wrote:
From: Andy Seaborne <[email protected]> Subject: Re: Fw: Re: [jena-dev]
Jena SDB: Nested transactions To: [email protected] Date:
Wednesday, 26 September, 2012, 1:06 PM
On 25/09/12 12:46, Suneetha wrote:
Hello,
This is related to the post: Jena SDB: Nested Transactions.
The database we are using is IBM DB2 database. Can we realize a
nested transaction monitor with this database? How difficult or
easy is it?
Why do you want to do this?
What does the "Type" column in the "NODES" table in the SDB store
indicate? We are not able to decode the number value in that
column.
For all practical purposes it is unused. It was to record the fact
it might be a known type (e.g. xsd:integer) but all info is in the
datatype.
Note that SDB uses a fixed schema for storing RDF and app does not
access the database directly but through the Jena API.
TDB is faster and scales better.
Andy
