Treatment of problems at the value level

Data format conversion, key conversion, constants, etc.

Challenges

There are many lower-level, data value-related problems that ETL architects face, such as:
 

Problem

Examples or details

Applying format masks to display or reinterpret values

ID'ing gender, ‚F‘ or ‚2‘ or mask date: mm/dd/yy to
dd-mm-yyyy

Homogenization of inconsistent production keys

Assignment of unique numerical substitutes for recognition and faster indexing

Pivoting and unpivoting

Transpose columns to rows and vice versa without duplicates

Inserting constants or literal character strings

Replacing or adding replacement values to protect or pre/re-label data

Setting or labeling values

Evaluate a condition to determine a standard label such as ‘NULL'.

Validation of values

Determine whether the input field is printable or has a specific data type.

Reformulation of values for adjustment to reference standards

Standardization of different data representations (e.g. country names)

Slowly changing dimensions in source tables (SCD)

Dissemination of new values (e.g. price change) in stock tables

Challenges

There are many lower-level, data value-related problems that ETL architects face, such as:
 

Problem

Examples or details

Applying format masks to display or reinterpret values

ID'ing gender, ‚F‘ or ‚2‘ or mask date: mm/dd/yy to
dd-mm-yyyy

Homogenization of inconsistent production keys

Assignment of unique numerical substitutes for recognition and faster indexing

Pivoting and unpivoting

Transpose columns to rows and vice versa without duplicates

Inserting constants or literal character strings

Replacing or adding replacement values to protect or pre/re-label data

Setting or labeling values

Evaluate a condition to determine a standard label such as ‘NULL'.

Validation of values

Determine whether the input field is printable or has a specific data type.

Reformulation of values for adjustment to reference standards

Standardization of different data representations (e.g. country names)

Slowly changing dimensions in source tables (SCD)

Dissemination of new values (e.g. price change) in stock tables

Solutions

The ETL/data management platform IRI Voracity and the IRI CoSort-tool that powers Voracity solve all of the above data processing problems and beyond:

  Keys and other data conversions
    Data type and format conversions
    Simple and advanced report generation
  PII masking and user-defined synthesis of test data

with a single 4GL and a data manipulation program called SortCL, which is also available in the Eclipse GUI called IRI Workbench is supported.

With SortCL you can:

    Reassign data and other data formats with new composite values and reusable data types
    Reassignment of source key names to comply with new ontologies
    Pivot (denormalize) for efficient queries and unpivot (normalize) before loading data into the warehouse
    Use inline ‘data' statements to add strings at any position and with user-defined repetition numbers
    Use conditional data or field logic to indicate qualifying values such as inserts vs. deletes vs. updates, etc.
    Use the provided data validation functions or third-party libraries to evaluate data characteristics or quality
    Use the integrated Perl Compatible Regular Expression (PCRE) logic to match and reassign data patterns
    Edit and report on multiple types of slowly changing dimensions and other changed data

IRI software such as CoSort and RowGen (test data generation) help you to generate unique initial or replacement keys, with functions such as:

    1. SEQUENCER, a specially named index field with user-defined start and increment values.
    2. random data generation to create random numbers or character strings that can also be sorted and decoupled.
    3. ROWID, a field with any name but specified as this attribute with characteristics outside SEQUENCER.
    4. UUID (and GUID), a transformation/generation function at field level that can be used in any part of a job.