Overview

When you are creating models, there are common problems scenarios that occur. You can create bits of models, called  "patterns" , which areto use in these scenarios.  You can then re-use the model patterns whenever the scenario occurs. Patterns can be:

• controls
• groups of controls
• portions of models

that solve problems that recur in the solutions that PhixFlow implements. Put another way, patterns are the bits of Models that are constantly re-used to solve problems that keep cropping up. There is no hard and fast rule for what constitutes a pattern, and some of the patterns listed here and in PhixFlow documentation and examples might seem very simple. There is also no such things as a definitive pattern list - new patterns can be added at any time. Employing these patterns and examples into PhixFlow models should help speed up model creation and development.

Common Scenarios

Data Collection and Comparison

• Removing duplicates
• Comparing two files
• Comparing Expected Vs Actual filenames in a directory
• Creating web service collectors
• Binary file loading
• Inexact / Fuzzy matching

Using Multipliers

• Turning a single record into multiple records
• Using Input Multipliers to insert values into the SQL Query of a DB Extractor
• Using Input Multipliers to break SQL "IN" clause into batches.
• Creating Unique indexes in PhixFlow
• Using InMemory Streams

 PhixFlow Operational Features

• Process Separate files into separate streams
• Using Concurrent Collectors
• Using Drill Down filters in lookups
• SQL Server Stream Sizing
• Oracle Server Stream Sizing

Other Patterns

• Data Enrichment Options : Loopback Lookup/Triangle pattern/Embedding fixed values in the query/merging directly from the DB.
• Managing a "Master List"
• General Ledger Reconciliation.

A pattern is any solution that is useful and reusable, from very simple methods to more sophisticated model-fragments.

This topic contains pattern suggestions for some common modelling scenarios.