Records

The Records tab shows each instance of an identified entity at the document level. You can review these Records in order to understand how normalization consolidated your Records into your final Entity List.

A Record represents every instance where the system finds an individual in the document set before normalization. If the system finds the same person in five documents, you will see five Records for that person. After normalization, the system creates a single Entity that represents that individual.

Review Records

To review documents in the Records tab:

Navigate to Entity Analysis and open the Records tab.
The Records list allows you to access all records included in the project and view details about each record.
You can customize details displayed in the Records list by clicking the Columns button to sort, filter, and rearrange the columns you want to include.
You can filter results in the Records list using the Filter function above each column. Type in a value to filter the results in a column.
To view advanced filter options, click into a column Filter box and select the Advanced button. You can use advanced options to enter a value and search using operators and other conditions.
Add the PI Value or PI Count columns to view PI information.
Review and validate that PI or sensitive data was correctly identified.
Confirm the context and accuracy of entity tagging such as names, emails, and SSNs.

Workflow example

Records are intended to assist users in QCing their final Entity Lists.

In order to view the Records for a large workspace, you will need to apply filtering to at least one column in the list. The application will indicate in the UI when the list is too large to generate all at once. Review the scenario below for an example of this workflow.

Records fields

Following is a list of Records fields:

Field	Description
Record ID	Unique identifier for the Record
Full Name	Full name including First, Middle, Last, Prefixes, and Suffixes. If Full Name is not identified as a single string in the original document, this field consists of the other fields combined.
First Name	First Name as entered on the document associated with this Record
Last Name	Last Name as entered on the document associated with this Record
PI Count	Count of pieces of PI that are linked to this Record
Control Number	The unique identifier for the document associated with the Record Records by definition are only associated with a single document, where an Entity can be associated with one or more documents.
Entity ID	The unique Entity identifier associated with this Record If normalizer has not been run since this record was created, there will not be an Entity ID value.
PI Types	All of the additional PI associated with the Records appear in the following columns
Annotation Types	A comma separated list of all PI Types associated with this Record
User Created	True/False field indicating if the Record was created by a human user or was system generated by Data Analysis
Status	This indicates if the Record is considered Valid or Invalid. Invalid Records will not be used in Normalization and do not appear in the Entity List.
Invalid Reason	The reason the Record was deemed invalid. This field is only populated if the Record is Invalid. SeeInvalid Reasons for details.
PI Columns	PI columns for all PI Types in use in your workspace.

Invalid Reasons

This feature is designed to exclude junk annotations from your Entity List. The vast majority of Invalid Records require no action and are correctly being excluded from the Entity List.

If you disagree with a Record being invalid, the table below will help you determine how to make the Record Valid. If you need to contact support for any reason regarding Records, including the applicable messages will improve your support experience.

Invalid Reason	Secondary Message	Description	Recommended Resolution
NAME_PARSING_FAILED	Can’t flip around multiple ‘,,’ characters in namestring.	The system is designed to flip “Last, First” names into “First Last” for formatting in the entity list. However, when more than one comma is present it cannot action.	To correct this you need to remove the excess commas from the name in the Record.
	Couldn’t find a last name in ‘{john doe}‘.	There is no last name identifiable in processing this Record. A standalone first name is not generally reportable data and is not enough to put through Normalization.	We recommend that you check the document for an appropriate Last Name for the Record. If one is not available, you can also put in a placeholder such as “Unknown”.
	Couldn’t find a first name in ‘{john doe}’.	There is no first name identifiable in processing this Record. A standalone last name is not generally reportable data and is not enough to put through Normalization.	We recommend that you check the document for an appropriate First Name for the Record. If one is not available, you can also put in a placeholder such as “Unknown”.
	Name could not be parsed by CJK or Latin parser	We attempt to parse names across multiple formatting options, however some names cannot be handled by the CJK (Chinese, Japanese, Korean) parser or the Latin parser.	We recommend that you review the documents containing these records and ensure the data has been properly extracted. If you still encounter issues, please contact support.
INVALID_NAME_ANNOTATION	Name string is empty	The system could not find a value in the name field (field is blank).	To correct this you need to add a valid name to the Record.
	Name string is ‘null’	The name for this Record is the exact string “null”, which we do not recognize as a legal name.	To correct this you need to add a valid name to the Record.
	Name has no tokens	The name field only contains non-alphanumeric characters (&*^%), which does not constitute a valid name.	To correct this you need to add a valid name to the Record.
	Name has only one token	We only support names with more than one token. This is to avoid junk names.	To correct this you need to add a valid name to the Record.
	Name has more than the max token limit of 6	We only support up to six tokens in a single name string. This is to avoid junk names that are full sentences or other “dummy” values.	To correct this, we recommend looking at the Name on the document and seeing if there are portions that can be excluded (for example, there may be many middle names and you can adjust the record to just first and last name).
	Name contains too many numeric characters	We support up to a single numerical digit in a name string.	To correct this you need to remove the excess numbers from the Name in the Record.
	Name contains too many punctuation characters	We support up to three non-alphanumeric characters in a name string.	To correct this you need to remove the excess punctuation characters from the Name in the Record.
REGEX_TIMEOUT	n/a	One of our validation tools in Normalization involves a regex validator.	This error is exceedingly rare - if you disagree with this designation please contact support.
MISSING_FULL_NAME	n/a	This can only occur if there has been a technical backend error.	If encountered, please contact support.
INACTIVE_FULL_NAME	n/a	This indicates that the name associated with this Record has been deleted from the UI.	To correct this, you would need to add the affected name back to the Record from the document viewer.
EXCEEDS_MAX_LINKED_ANNOTATIONS	n/a	We currently support up to 50 unique PI Annotations connected to a single Record. We have never seen this limit exceeded in a customer project.	If you exceed this limit, you can remove the least important PI links to get below 50 and it will run through Normalization. Please also contact us and we would love to hear your use case.

Entity information

You can view information about a Record, including the Entity it belongs to after normalization, by selecting the Record ID. This will take you to the Entity Information view.

If normalization has not been run since this Record was created, then you will not be able to view the Entity Information panel.
Entity information window

The Entity Information view matches the details you see when you select an Entity ID in the Entity Report.

When you open the Entity Information view from a Record ID, the system highlights that specific Record in yellow whenever the Entity contains multiple Records. This highlight helps you understand how the Record you selected fits into the full Entity.

Save as List

Select records from the document list and choose the Save as List mass action to generate a list of documents which contain the selected records.

This list can be helpful for managing your review, especially if you have identified key errors or inconsistencies in your Entity List.

Download

Select the Export to File button to save an .xslx or .csv file to your downloads folder. Any filtering or column changes applied to the report in the UI will be maintained in the downloaded file.