Striim 3.10.3 documentation

How update and delete operations are handled in writers

In some cases, when the input of a target is the output of a CDC source, Striim can match primary keys or other unique identifiers to replicate source update and delete operations in the target, keeping the source and target in sync.

In other cases, such matching is impossible, so updates and deletes are written as inserts. This is often acceptable when the target is a data warehouse.

With the following source-target combinations, update and delete operations in the source are handled as updates and deletes in the target:

source

target

  • GG Trail Reader

  • HPNonStop readers

  • MSSQL Reader

  • MySQL Reader

  • Oracle Reader

  • PostgreSQL Reader

  • Salesforce Reader

  • Azure Synapse Writer

  • BigQuery Writer (in MERGE mode)

  • CosmosDB Writer

  • Database Writer

  • Hazelcast Writer

  • HBase Writer

  • Kudu Writer

  • MapRDB Writer

  • MongoDB Writer

  • Redshift Writer

  • Snowflake Writer

  • Spanner Writer

  • MongoDB Reader

  • CosmosDB Writer

With the following source-target combinations, update and delete operations in the source are handled as inserts in the target:

source

target

  • GG Trail Reader

  • HPNonStop readers

  • MSSQL Reader

  • MySQL Reader

  • Oracle Reader

  • PostgreSQL Reader

  • Salesforce Reader

  • BigQueryWriter (in APPENDONLY mode)

  • Cloudera Hive Writer

  • DatabaseR eader

  • Incremental Batch Reader

  • Azure Synapse Writer

  • BigQuery Writer

  • CosmosDB Writer

  • Database Writer

  • Hazelcast Writer

  • HBase Writer

  • Kudu Writer

  • MapRDB Writer

  • Redshift Writer

  • Snowflake Writer

  • Spanner Writer

When the target is HiveWriter or HortonworksHiveWriter, update and delete operations may be handled by the target as updates and deletes or as inserts depending on various factors (see Hive Writer for details).Hive Writer