Release Version 2.28.3

In this release, we have addressed the following issues to improve the usage and functionality of our product features. To know the list of features and integrations we are working on next, read our Upcoming Features page!

The content on this site may have changed or moved since you last viewed it. As a result, some of your bookmarks may become obsolete. Therefore, we recommend accessing the latest content via the Hevo Docs website.

In this Release

• Fixes and Improvements

Fixes and Improvements

Destinations

• Handling of Reserved Keyword Column Names in Azure Synapse Analytics Destinations

  • Fixed an issue where Hevo was unable to load data from tables with column names that were the same as reserved keywords in Azure Synapse Analytics. This issue occurred because these column names were mistakenly treated as keywords in the query that loaded data into the Destination, causing syntax errors. With this fix, reserved keyword column names are now enclosed in double quotes in the query, preventing them from being treated as keywords and allowing for successful data loading.

Sources

• Handling Data Ingestion for Materialized Views in PostgreSQL Source

  • Fixed an issue where materialized views were being ingested during incremental loads in a log-based Pipeline created with any variant of PostgreSQL, even though Hevo does not support data replication from views. This issue occurred because Hevo mistakenly identified materialized views as regular tables during incremental loads, resulting in higher data ingestion and costs.

    With this fix, Hevo now checks if a table is a materialized view during incremental loads and skips it, ensuring only supported tables are ingested. This update will be deployed to teams across all regions in a phased manner and does not require any action from you. However, to enable this fix for your existing or new Pipelines, contact Hevo Support.

• Handling Incorrect Alerts for Deleted Folders in Google Drive Source

  • Fixed an issue where users received incorrect alerts after deleting a folder in Google Drive. When a folder is deleted, Hevo marks the object and task statuses as finished to indicate its removal. However, the following issues occurred:

    • If the Pipeline was not updated, the statuses were not marked as finished, inaccurately indicating them as still present.

    • If the Pipeline was updated, only the object status was marked as finished, leaving the task status unchanged.

    As a result, when Hevo attempted to ingest data from Google Drive afterward, it checked the statuses of all folders. Because the statuses of the deleted folders were not updated correctly, Hevo indicated that these folders still existed, which triggered alerts to users about missing folders. With this fix, both the object and task statuses are now accurately marked as finished when a folder is deleted, preventing incorrect alerts.

• Handling of Fulfillment Orders Object in Shopify Source

  • Improved Data Ingestion: Increased the record limit for the Fulfillment Orders object from 50 to 250 in the Shopify integration. This change allows for more data to be ingested from the object, reducing the chances of records being missed.

  • Data Mismatch Issues: Fixed an issue where the API call to fetch the Fulfillment Orders object in Shopify returned incomplete data. This issue occurred because the fulfillment_holds field was not included while parsing the API response. With this fix, the system now includes this field during parsing, ensuring that all data from the object is fetched.

    This fix applies to Pipelines created with Shopify as a Source after Release 2.28.3. If you observe a data mismatch for the Fulfillment Orders object in your existing Pipelines, you must create a new Pipeline.

• Improved Offset Management in Klaviyo v2 Source

  • Fixed an issue in the Klaviyo v2 integration where the API call to fetch the Events_v2 object returned incomplete data. The polling job used the timestamp of the last record retrieved as the offset for the next poll. However, the records were not sorted by timestamp, and the next poll fetched data from a timestamp later than the offset. As a result, data between the previous and new offsets was not ingested. With this fix, Hevo now sorts the records by their timestamps, ensuring that all data is fetched in the correct order and preventing data loss.