Pre-Conversion Tasks

Performing Data Integrity Analysis

The Data Integrity (DI) report is the cornerstone of Gladstone’s methodology. It is the one piece of analysis that has proven to be the most valuable when performing a data conversion. Data to be converted consists of data (values) in fields, fields in records, and files (databases) that are used by applications. The Gladstone analysis shows you all the values that reside in every field of each data file. The rule is never to move a field’s contents unless you know for sure what is in the field. The DI report is useful to programmers, business analysts, managers, and auditors — anyone who is interested in the exact contents of the data. The truth is in the data!

Converting File Definitions to Common Base

Gladstone provides several programs (called filters) that convert non-COBOL file definitions (for example, Assembler, PL1, and RPG) to COBOL file definitions. This allows the easy transformation of non-COBOL file definitions to COBOL file definitions, which the Gladstone software uses exclusively as a common base for the total conversion suite.

Finding Dates in Data and Source Code

Many software products locate dates by searching source code for field names that imply that the field may contain a date. Looking for names such as “date” or “year” is neither efficient nor effective because date fields can have any label. Searching the source code is at best a hit-and-miss method. The Gladstone approach is to search the actual data for date formats. For example, even if the field name is “WXYZ-GEORGE” the Gladstone software determines that the format is CCYYMMDD and lists the field as a potential date field. After finding all potential date fields, the Gladstone software creates a seed file that lists the actual field names of the date fields. The seed file is fed to other Gladstone software to scan the source code for dates. A Gladstone program uses the field names found in the analysis of actual client data and scans the source code for every place that the field name is referenced. The software produces a report that shows the field names underlined and displays the five lines of code before and after the occurance. Another report shows how many times each field name is found.

Processing “Foreign” Data

Gladstone provides several sets of programs that convert “foreign” data and file definitions into IBM data and file definitions. This allows data that is not IBM compatible to be moved to the IBM platform with ease (Unisys to IBM, NCR to IBM, Data General to IBM, ASCII to EBCDIC.) The software generates programs that convert the data from a foreign format to an IBM-compatible format.

Charting Client’s Old to New Products

The Product Mapping report is used by clients during certain types of conversions such as corporate acquisitions and mergers. Such conversions frequently involve the merging of product lines. For example, if a bank acquires another institution, the product offerings such as checking and savings account options must be synchronized. Traditionally management immediately has sent representatives to the acquired entity in order to initiate the mapping process. With the use of the Gladstone software, the team now analyzes the data for about one week prior to the site visit, and then goes out to map products. At that point, the team usually knows the data better than the acquired entity. This process eliminates multiple visits, confusion, discovery of “grandfathered” products, etc. This has proven to be a standard process for the clients of Gladstone software in the banking field. it is applicable to other industries and situations.

In a typical conversion many data elements end up not being used. Also, it is difficult to determine exactly where each element is used and how many times it is used. The Gladstone software creates a set of reports that shows the elements that are used, elements that are not used, for elements used, where each element is used. This is performed for each source file used to generate the target files.

Testing is a major function in any conversion. In order to test, clients must have test data. they can use the full file, the first several records, or selected data. Testing with the full file is often not practical, and any subset of data may not be truly representative. Planning a course of action is time consuming and often does not produce an adequate base of test data. The Gladstone Test Data Generator uses parameters entered by the client to generate the test data set with a minimum of effort.

The process involves running the Data Integrity program, which creates a report that shows all the different values for each element (up to 20 by default.) In almost all cases where there are more than 20 values for an element, the element does not have any unique type of testing criteria. The Data Integrity analysis indicates the key of the first record that was found with each unique value and writes a file containing all of these keys. The keys are sorted, and duplicates are eliminated. The output of this process is then used as input to a program that reads the original input data file and writes out only those records for which a key has been created. The client has the option to add keys of specific records to the test set. The test file is one of manageable size (usually 1200 records or less.)

For security reasons, it is sometimes not appropriate to test with even a subset of live production data (right of privacy issues, HIPAA, etc.) Gladstone software can generate a set of fictitious or “alias” data that substitutes for the real contents of designated fields such as names, account numbers, or unique identification numbers to protect confidential information. To generate alias test data, the Gladstone software takes actual data records and file definitions and substitutes alternate contents.

A key to Gladstone’s power lies in the methodology used whereby the actual conversion code is generated dynamically based on the specifics of the data migration at hand. This provides extreme flexibility to handle an unlimited array of different conversion situations.