Profiler

JBerd Profiler

Overview

The purpose of the profiler is to provide a report on the values and frequency of occurrence of fields and structures in an input file.

This enables us to learn:

• The observed values in data fields
• Which fields are unique within a file
• Which fields have limited value sets, and what those values are
• Which fields are optional
• Which fields are mandatory
• Which combinations of optional fields occur within records
• The observed count of repeating items

Consider, for example, a BER encoded file that contains many records, called “Xrecords” and that each Xrecord may contain fields A, B, and C, of which A and B are mandatory C is not.
The profiler will show:

** At the record level ** The total number of records seen, and the number of records with each combination of fields. So if there were 100 records, all of which contained A and B but only 10% contained C, we would see the counts of the different combinations as follows:

• A,B,C 10 (only 10% contain C)
• A,B 90 (90% contain only A and B)

** At the field level ** The total number of instances of each field, the total number of distinct values, and the first 100 distinct values. So if, in our example, A was a unique id then we would see there would have been 100 instances and 100 distinct values. But if C was a Y/N flag, and only one of the instances was “N” we would see 10 values, 2 distinct values, and a count of each distinct value: Y(9); N(1).

Field names and values

The output of the profiler identifies fields and records by name, provided that such names have been defined in an input alias file. See [Tags and Aliases]

The way in which the field values is printed (rendered) is governed by the default rendering method provided in the alias file. See4 [Rendering]

Further Information

To see how alias names and types are associated with BER data items see [Tags and Aliases]
To see how the binary BER data is rendered as printable output see [Rendering]

For full documentation see Jberd.pdf