Sybase 12.4.2 User Manual

Page of 536
Importing data by replication
226
 
Space for deletions
When you use the 
DELETE
 statement, you may need to add space to your 
database, due to the way Adaptive Server IQ stores versions of data pages. For 
details, see “Overlapping versions and deletions”.
When you use 
DROP TABLE
 or 
TRUNCATE TABLE
, you do not need to add 
space, as no extra version pages are needed.
Importing data by replication
If you need to update your IQ data frequently from an Adaptive Server 
Enterprise database, you may want to consider setting up a replication 
environment. In this environment, you can use the Sybase Distribution 
Director to automate the process of replicating data from an Adaptive Server 
Enterprise database into an Adaptive Server IQ database.
Distribution Director establishes a two-phase process in which transactions 
that occur on the source database—the Adaptive Server Enterprise database—
are repeated on the Adaptive Server IQ database. To make this process work, 
you need:
An operating Replication Server system, which includes a special database 
called the Replication Server System Database (RSSD). This database 
contains replication definitions that describe the transactions in the source 
database, and how they should be replicated to the replicate databases.
A specially prepared replicate database on an Adaptive Server Enterprise 
database. This replicate database serves as a staging area for the migration 
into your IQ database.
Your IQ database set up with the appropriate tables for the data to be 
replicated.
Here is a summary of how it works:
1
In Phase 1, the Replication Server replicates transactions from the source 
database to the RSSD on a continuous basis.
2
In Phase 2, the Adaptive Server IQ staging migration is executed, to 
transfer the replicated transactions from the RSSD into the IQ database.
This two-phase approach overcomes differences between the way Adaptive 
Server Enterprise and Adaptive Server IQ store and update data. It also allows 
you to schedule insertions into your database for lower usage periods, and 
avoids conflicts with other insertions or loads to the same tables.