Recent changes to Profiler

Profiler modified by Andrew Forsyth

Andrew Forsyth — Thu, 21 May 2015 16:51:36 -0000

--- v4
+++ v5
@@ -27,6 +27,13 @@
 ** 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]

Profiler modified by Andrew Forsyth

Andrew Forsyth — Thu, 21 May 2015 16:48:51 -0000

--- v3
+++ v4
@@ -25,3 +25,14 @@
 * 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).
+
+
+### 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
+

Profiler modified by Andrew Forsyth

Andrew Forsyth — Thu, 21 May 2015 16:47:18 -0000

--- v2
+++ v3
@@ -5,6 +5,7 @@
 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

Profiler modified by Andrew Forsyth

Andrew Forsyth — Thu, 21 May 2015 16:46:48 -0000

--- v1
+++ v2
@@ -1 +1,26 @@
 ## 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).

Profiler modified by Andrew Forsyth

Andrew Forsyth — Thu, 21 May 2015 16:43:05 -0000