We are excited to release a major new integration with the cloud data warehouse service - Snowflake! Sign up to try it out today, or, if you're an existing Alooma user and want to try it out, just contact us.
Other than Snowflake, we're continuing to smooth some rough edges of our product. Read on to learn more.Auto-mapper Improvements
Adjusting the auto-mapper to work with various types of data warehouses is a long process which we've been working hard on. Abiding by naming constraints and conventions, type conversions, and much more can be very different between BigQuery, MemSQL, Snowflake, and Redshift. So far we've made the basic adjustments to allow the auto-mapper to work, but there's still much work to be done to make it awesomely accurate in it's schema translations.
We've also fixed several auto-mapper bugs:
- When mapping new fields of an event type, the mapper will use matching existing columns (instead of always creating new ones).
- Not many know this, but when translating a MySQL VARCHAR to a Redshift VARCHAR, you need to multiply its length by 4. We knew that, but forgot to make sure we don't pass the Redshift VARCHAR length limit of 65535. Well, now we take care of that too.
- When using the auto-mapper button from the Mapper screen with a table selected, the auto-mapper will no longer fail.
- We fixed a broken schema translation in our Salesforce input (metadata_deleted was wrongly mapped to VARCHAR)
- Fixed our own metadata mapping.
- We added quoting to the tables and columns the auto-mapper creates, so we also removed the conversions it used to do (dashes to underscores, camel case to snake case, etc.). So, even though Redshift's table and column names are case insensitive, Alooma's auto-mapper will make the best effort to keep them as close to the source as possible.
A release would not be a release without polishing the UI:
- We fixed a false alert indicating that the data warehouse configuration is missing.
- The Plumbing screen's loader will no longer linger on the screen, after the Plumbing has loaded.
- A bug in grouping different Dashboard notifications to a single notification has been fixed.
- We improved how the log in screen looks on mobile (but full mobile experience is not there yet).
- We changed the position of the Purge Restream button in the Restream dialog to a less confusing location.
- We improved our MongoDB input to be more resilient and communicative in the face of network errors or permissions misconfiguration.
- The ODBC-based inputs (MSSQL, Oracle, etc) now support primary keys with upper case letters, and specifically we fixed our MSSQL connector to support update indicator columns of type BIGINT, and added the support of IDENTITY columns.
We are hard at work at taking our post-loading consolidation to the next level. We plan to expose it to our users and allow you to configure scheduled queries to incorporate post processing into your plumbing. We use it internally to consolidate change logs to actual tables, but you can use it for much more (expect a blog post when we fully release it!).
We just made the following improvements to our query scheduler:
- Support for SSH tunnels has been improved.
- We improved the integration of the Mapper and query scheduler.
- We've made the necessary preparations to expose change log consolidation configuration in the UI, and via alooma-python.
Interested in learning more? Leave us a comment, or send us an email!