To load a big table in Qlik Replicate follow these steps:
- Create a view for each partition on the source:
select 'CREATE OR REPLACE VIEW ' || table_owner || '.' || 'SALES_VIEWS_' || PARTITION_NAME ||
' AS SELECT * FROM ' || table_owner || '.' || TABLE_NAME || ' PARTITION(' || PARTITION_NAME || ');'
from dba_tab_partitions
where table_owner = 'SH'
AND TABLE_NAME = 'SALES';
- Add the table to the existing task and set error handling to upsert. The table will start getting the changes.
- Create a new task (full load only) and just select the views that were created. This will create table on target with same name as view. Then crank up the number of tables loading in parallel until it squeals.
- As each view loads merge the corresponding view_table on the target to the base table. They could use a merge syntax that would only copy rows that do not exist.