I85PREP
⌘ K
Informatica Data Engineering Foundation study materials
Informatica Data Engineering Foundation study resources

Informatica Data Engineering Foundation: Data Integration

Informatica Data Engineering Foundation · Work through your exam, one topic at a time.

All guide sections

Data Integration

Data Integration - what the exam tests

The Data Integration domain checks whether you can read a mapping's logic and predict its runtime behaviour: which transformations change row counts, how parameters resolve, and where pushdown or partitioning alters the result. Expect mapping-design questions, session/task configuration, and error-handling scenarios framed in both PowerCenter and Cloud Data Integration terms.

Core rules

  • A mapping is a logical plan (sources → transformations → targets); a session or task executes it, and a workflow or job sequences tasks.
  • Active transformations change row counts: Filter, Router, Aggregator, Rank, Joiner, Union, Update Strategy, Sorter with Distinct, Source Qualifier with a filter. Passive ones do not: Expression, Lookup, Sequence Generator, Stored Procedure.
  • Mapping parameters resolve at run time from a parameter file; mapping variables can be reassigned inside the mapping. In Cloud Data Integration, in-out parameters pass values between parent and child tasks.
  • Pushdown optimization moves logic into the database (source, target, or full). Non-pushdown-compatible transformations force a fallback or a failure depending on configuration.
  • Update Strategy drives target behaviour: DD_INSERT, DD_UPDATE, DD_DELETE, DD_REJECT.

Worked example

A mapping loads a customer dimension as SCD Type 2. A Lookup on the dimension returns the current active row; an Expression compares the incoming hash of tracked columns against the stored hash and sets a changed flag; a Router sends changed rows to an Update Strategy (DD_UPDATE) and new rows to DD_INSERT; unchanged rows fall through a Filter that drops them. Reasoning: Lookup is passive, so it cannot remove rows - the Router and Filter do that work. If the lookup condition matches several rows, the session fails unless the lookup is explicitly configured to return any or all rows.

Traps

  • Assuming Lookup filters rows; it only returns ports.
  • Forgetting Sequence Generator reset in partitioned sessions, producing duplicate keys.
  • Treating a mapplet as a reusable transformation: a mapplet contains a flow, a transformation does not.