Transact

⌘K
  1. Home
  2. Transact
  3. Release Notes and Softwar...
  4. 2020.1.02
  5. Known Issues and Workarounds 2020.1.02

Known Issues and Workarounds 2020.1.02

The following is a list of known issues for which a workaround is provided.

Category Known Issue Workaround
Batch Class Import If you are importing a batch class that creates or updates global document types in the global batch class, the import process will fail if there is a local document type in the Global Batch Class (BC0). Importing will succeed if all local document types are changed to global for BC0. Perform the following as a workaround:

  1. In BC0, set any existing local or draft document types to Global.
  2. Import your batch class.

Note: If you do not want to keep your local or draft document types as Global, perform steps 3 and 4.

  1. Create copies of the document types you set to Global in step 1.
  2. Delete the original document types you set to Global in step 1.
Batch Class Import/Export When exporting a batch class that has Batch Class Fields, importing that batch class does not import the Batch Class Fields. To resolve this issue, manually add the Batch Class Fields after import.
Chrome 80 Due to new restrictions in cross-site cookies with Chrome 80, you may have issues with interoperability with iframe and third-party web apps. To resolve this issue, you may need to adjust your cookies. Refer to Additional Configuration for Chrome 80 for steps.
ICR/OMR Extraction When processing skewed or rotated images using ICR-related extraction methods, characters are not extracted properly, and any noise on the page may cause random characters to be extracted. To resolve this issue, manually deskew or rotate the images and send them in as a new batch instance.
ICR/OMR Extraction Repeatedly cycling between Test Extraction and Back causes the temp files to grow quickly for the batch class. To resolve this issue, manually remove temp files as needed.

  1. Go to [Ephesoft_Directory]\SharedFolders\[batch class]\test-advanced-extraction/snippets
  2. Delete test files from this folder as needed.
ID Extraction When performing review or validation of fields extracted using ID Extraction, some batch instances may get stuck in Ready for Validation despite extracting the correct data. To resolve this issue, set the OCR Confidence Threshold to zero (0).

  1. Go to the Index Fields page for the batch class.
  2. Locate the index fields that are turning red on the validation screen.
  3. Set the OCR Confidence Threshold for these index fields to zero (0).
ID Extraction When extracting data from ID types with the date format MM/DD/YYYY using the ID Extraction plugin, the return results are in the format DD/MM/YYYY. To resolve this issue, either create a format conversion rule or custom script to change the date field back to the desired format.

For more on format conversion, refer to this help article.

ID Extraction The following licensing error may occur in the dcma log after installing 2020.1.02:

“ID Extraction License isn’t enabled or expired”

If you are not using the ID Extraction plugin, no action is required.

If you would like to use the ID extraction feature, you will need a new license. Contact tickets@ephesoft.com with your request.

ID Extraction Some ID types may not successfully extract the back of the ID. For example, when extracting data from the back of a Mexico ID card classified as a MEXICO_ID document type, index fields may be missing or the batch instance may go into error. To prevent this issue, classify the back of an ID card using the GLOBAL_MRZ document type.
ID Extraction When using a license that is not specific to 2020.1.02, adding the IDENTIFICATION_EXTRACTION plugin to a batch class and setting it to ON will cause any batch instances run in this batch class to stop indefinitely at Waiting to Restart status instead of receiving an error message explaining that you do not have the appropriate license.. If you want to use the IDENTIFICATION_EXTRACTION plugin, contact your sales representative to purchase the ID Extraction feature. You will need to install a new license with ID Extraction enabled.
Index Field Import When importing an index field, the Default Value Batch Class Field fails to import. This issue does not occur when copying an index field. To prevent this issue, note the Default Value Batch Class Field before export and manually add them to the index field after import.
Installation When performing a fresh installation of 2020.1.02, you may be unable to install the Ephesoft LDAP service.

This is due to the Ephesoft LDAP having a dependency on Microsoft Visual C++ 2010 Redistributable.

If your security needs permit it, install Microsoft Visual C++ 2010 to resolve this issue.

Otherwise, you will need to install the newest version of OpenLDAP separately from Transact, which does not have the dependency on VCRT 2010. Refer to OpenLDAP: Separate Installation for instructions.

KV Extraction If you switch between extraction methods in the KV Extraction screen, the Minimum Confidence Threshold field disappears when Hand/Machine Print is selected as the extraction method. To resolve this issue, perform either of the following:

  • Refresh the page.
  • Toggle between extraction types (Go from KV + Hand/Machine Print to Hand/Machine Print and back to KV + Hand/Machine Print).
KV Extraction When testing KV Extraction using Hand/Machine Print as the extraction method, if the key regex matches multiple results on the page, the value snippet will not extract from the correct area. To prevent this issue, choose a unique key regex pattern when defining a KV extraction rule.

For example, if the word “Number” shows up multiple times on a page, do not set the regex to only “Number”. Instead, use the full field name. Some key regexes could be:

  • Social Security Number
  • File Number
  • Insurance Number\(s\)
KV Extraction When editing an extraction rule on the KV Extraction screen, saving and re-editing causes an error loading the image file. This error occurs when Ephesoft Transact picks an orphaned file that doesn’t exist in the backend.

To resolve this issue, select the correct file from the document dropdown on the KV Extraction screen.

KV Extraction When Auto-Resize K/V Areas is enabled on a KV extraction rule for normal documents (unskewed, not rotated), the extracted data is less accurate compared to KV extraction rules with the option disabled. To prevent this issue, uncheck Auto-Resize KV Areas.

Otherwise, when drawing the key overlay, ensure there is no extra margin around the key text. This may require testing and refining the overlay size.

Known Issues

The following is a list of known issues for which there is not yet a workaround or fix.

Category Known Issue
Email Configuration After setting up an email import in Email Configuration, clicking Apply causes the Client ID and Client Secret values to display as a hash instead of being masked.
ICR/OMR Extraction For Hand/Machine Print extraction, selecting Numeric as the Value Type still allows for alphabetic results to be returned. This is scheduled to be resolved in 2020.1.03.
ICR/OMR Extraction When Checkbox Detection is selected as the extraction method and the Pixel Density % is configured, toggling to Hand/Machine Print, editing the Minimum KV Confidence, and toggling back to Checkbox Detection causes the Pixel Density % to change.
ID Extraction On the Test Extraction screen, the IDENTIFICATION_EXTRACTION plugin is missing from the Extraction plugins field despite being deployed in the workflow.
ID Extraction When using a license where ID Extraction is set to OFF, you are still able to add the IDENTIFICATION_EXTRACTION plugin to the Extraction module, however you won’t be able to use the plugin.
ID Extraction If the IDENTIFICATION_EXTRACTION plugin fails to extract a field when other plugins are deployed in the Extraction module, the extraction type will change to one of those other plugins.
ID Extraction When adding the IDENTIFICATION_EXTRACTION plugin to a batch class for the first time, the plugin is correctly set to OFF by default. However when you add it to a second or subsequent batch class, the plugin is ON by default.
ID Extraction If a document is incorrectly classified in the review stage for the ID Extraction batch class, irrelevant data may be extracted in some fields.
ID Extraction Licensing After upgrading from 2020.1 to 2020.1.02 with an old license, under Licensing Details, the image count for ID extraction incorrectly displays as “Unlimited” instead of “N/A”.

Despite this, you will be unable to process ID extraction batch classes. If you have not purchased the ID extraction feature, no action is required.

If you would like to purchase the ID extraction feature, you will need a new license. Contact your local sales representative with your request.

Installation When upgrading and launching the installer for the first time, Windows may report the installer as Not Responding on the End User License Agreement screen. Do not cancel or kill the installation, it will recover and resume (~5 minutes).
KV Extraction On Linux (RHEL), the following fields are missing from the left panel on the Paragraph Extraction Rule page:

  • Start Pattern
  • End Pattern
Licensing On the License Details page, the License Expiry Date is displayed in Greenwich Mean Time (GMT) while the Identification Extraction Expiry Date is displayed in Coordinated Universal Time (UTC).
Logging An error occurs in the dcma logs when exporting a batch class, even though the batch class exports successfully.

Error: “unable to export CHE plugin configuration file while exporting BC5org.springframework.web.client.HttpClientErrorException: 404 null”

We encountered this error during our testing. This had no impact on the performance or data that was returned.

Logging When processing batches in a multi-server environment, the following errors may occur in the dcma-all log:

“Error copying the serialized file” and “Unable to get workflow details for batch instance”

We encountered this error during our endurance testing. This had no impact on the performance or data that was returned.

Logging When using Windows 2016 and Oracle 18c, the following error appears in the dcma logs:

“ORA-02275: Such a referential constraint already exists in the table”

We encountered this error during our testing. This had no impact on the performance or data that was returned.

Logging When using a previously installed license without Identification Extraction, the following error regularly occurs in the logs:

“Error occurred while computing IDE remaining Image Count.”

This has no impact on batch processing.

Logging When using Microsoft SQL Server as your database, the purging job successfully runs to migrate batch instance data from the Report database to the Report_Archive database, however an error about duplicate data appears in the logs.
Logging On a multi-server environment, the following error may occur in the dcma-all.log:

“Server… not able to release lock over batch class ID BC6 for operation R for Machine Learning”

Further analysis in our testing scenario indicated no loss in performance or data.

Logging When Transact is installed on Windows, the OS name may display incorrectly in the dcma-all.log. For example, when installed on Windows Server 2019, the OS name displays as Windows Server 2016 in the dcma-all.log.

This has no impact on the functionality of Transact.

Logging In the Email Configuration page for a batch class, when the email import Protocol is set to OAuth2, entries are incorrectly recorded in the log files for each email import cron job. Rollover logging should prevent the file size from growing too large.
Logging When using Oracle Database 18c, you may observe errors in the logs regarding the alter table query.

We encountered this error during our testing. This had no impact on the performance or data that was returned.

Reporting The following Dashboard reports fail to export as PDFs:

  • Server Status
  • Pages Processed per Batch Class
  • Average Pages Processed

These reports can still be exported as Excel files.

Reporting Filters that are set in the Reports module do not persist when exported. Only the default data associated with the report is exported.
User Interface (UI) On the Plugin Configuration page for CREATEMULTIPAGE_FILES, the option ITEXT-SEARCHABLE is displayed twice in the Multipage File Export Process dropdown. The second option is not selectable, and this had no impact on any functionality during our testing.