1.5.1

Nothing has changed here, just ensuring that the deprecation notice appears on the package hub.

dbt_metrics is replaced by metricflow. For more information, please reference this dev blog post!

1.5 Release & Deprecation Warning

Deprecation Warning

This will be the final minor version release of dbt_metrics. For more information, please reference this dev blog post!

What's Changed

• Fixing Calendar YML by @callum-mcdata in #230
• Adding error message for derived metric filters by @callum-mcdata in #216
• Updating Metric Expression by @callum-mcdata in #238
• Adding support for 1.5 by @callum-mcdata in #240

Full Changelog: 1.4.1...1.5.0

1.4.1 - Bug Fixes

What's Changed

• Error Message For Ephemeral Models by @callum-mcdata in #222
• feat: SEMANTIC-154 - Add properties file for calendar model by @akv-akv in #221
• Update TOC to remove reference to all_time grain by @jthigpen in #220
• removing calendar tests by @callum-mcdata in #223

New Contributors

• @akv-akv made their first contribution in #221
• @jthigpen made their first contribution in #220

Full Changelog: 1.4.0...1.4.1

1.4.0 - Performance Improvements

New Features

• Even in the original issue for metrics, there were users suggesting that timestamps should be an optional property and that some metrics can't be time-bound. This was deprioritized during the initial implementation but with the release of dbt-core 1.4, we've made this a reality! Metric timestamps (and time_grains) are now an optional property of a metric and the metric macros can now calculate metrics without needing to provide a grain. grain is now an optional input to the metrics macros and if it is not present then the generated SQL will aggregate the metric without any consideration for time. PR for the curious
  • This piece of work changes the SQL generation to ensure that there won't be 0 values for dimensions that don't have values on the particular date. If you're familiar with the way we generate sql, it means we're no longer date-spining fields unless it is required by the query.
• Median has been the top requested calculation method for the past year. So we decided to sprinkle in some post-holiday presents and release it as part of 1.4! For more information on how we implemented this, especially for data platforms that don't have a median aggregation, see the PR.
• Adding date-alias input to macro in order to support BI tools that require statically named columns by @callum-mcdata in #209

New Supported Adapters

• With the introduction of dbt-databricks support in dbt Cloud, we wanted to make sure that dbt_metrics had parity! Databricks is now a supported adapter for dbt_metrics.

Bug Fixes/Small Improvements

• Making metrics queries more performant by looping through common groups instead of metrics @callum-mcdata in #211
• Code cleanup (re: issue #144) by @deanna-minnick in #189
• Reducing Order By Complexity by @callum-mcdata in #191
• Hotfixing misnamed variable by @callum-mcdata in #207

New Contributors

• The wonderful @atomatize made their first contribution with a documentation fix
• The documentation guru with eagle eyes @mirnawong1 spotted some mistakes and made fixes!

Full Changelog: 1.3.2...1.4.0rc1

1.4.0rc2 - Making Metric Queries Smarter

What's Changed

• Adding date-alias input to macro in order to support BI tools that require statically named columns by @callum-mcdata in #209
• Making metrics queries more performant by looping through common groups instead of metrics @callum-mcdata in #211

Full Changelog: 1.4.0rc1...1.4.0rc2

1.4.0rc1

New Features

• Even in the original issue for metrics, there were users suggesting that timestamps should be an optional property and that some metrics can't be time-bound. This was deprioritized during the initial implementation but with the release of dbt-core 1.4, we've made this a reality! Metric timestamps (and time_grains) are now an optional property of a metric and the metric macros can now calculate metrics without needing to provide a grain. grain is now an optional input to the metrics macros and if it is not present then the generated SQL will aggregate the metric without any consideration for time. PR for the curious
• Median has been the top requested calculation method for the past year. So we decided to sprinkle in some post-holiday presents and release it as part of 1.4! For more information on how we implemented this, especially for data platforms that don't have a median aggregation, see the PR.

New Supported Adapters

• With the introduction of dbt-databricks support in dbt Cloud, we wanted to make sure that dbt_metrics had parity! Databricks is now a supported adapter for dbt_metrics.

Bug Fixes/Small Improvements

• Code cleanup (re: issue #144) by @deanna-minnick in #189
• Reducing Order By Complexity by @callum-mcdata in #191
• Hotfixing misnamed variable by @callum-mcdata in #207

New Contributors

• The wonderful @atomatize made their first contribution with a documentation fix
• The documentation guru with eagle eyes @mirnawong1 spotted some mistakes and made fixes!

Full Changelog: 1.3.2...1.4.0rc1

1.3.2 - Bug Fixes and Rolling Secondary Calc Change

The 1.3.2 Release - Bug Fixes and Rolling Secondary Calc Change

This release is mainly bug fixes with a single changed feature that had been requested by @andresrecalde. We also had a new community contributor in the form of @rijnhardtkotze and @lianndellandrea! We love to see the community feedback in the form of issues and pull requests!

Changed Features

• We've altered the rolling secondary calculation to make the interval input optional. If no interval is provided then the secondary calculation becomes unbounded and does a unbounded aggregation on all the previous rows. Helpful for keeping a running total from perpetuity! Change by @callum-mcdata in #186. Thanks for the suggestion @andresrecalde!

Bug Fixes

• Fixed a bug where the calendar table was incorrectly joining to the base model when window was enabled on the metric. This was including an extra day in cumulative metrics. Change by @callum-mcdata in #183
• Fixed a bug where start/end date fields ran into namespacing issues with the calendar table. date_day in both caused a bad time! Change by @callum-mcdata in #185

Project Changes

• Fixes for JIRA Workflow by @callum-mcdata in #167
• Updating Jira Project for Actions by @leahwicz in #169
• Updating other Jira Actions to refer to the SEMANTIC project by @leahwicz in #171
• [Bug] metric_jsonschema_example still references old properties in required fields by @rijnhardtkotze in #177
• [SEMANTIC-65] [Bug] Environment variables in tests doesn't match test.env.example file by @rijnhardtkotze in #179

Full Changelog: 1.3.1...1.3.2

1.3.1 - Bug fixes from the community

What's Changed

Hey folks! Well, we're one week past Coalesce and that means life is getting back to normal - ie, time for some smaller changes and bugfixes!

Bug Fixes

• @JoeryV found a bug with window metrics where the day grain would cause the query to fail - and then submitted the fix himself!! A great example of a community member taking on the challenge of fixing the issues they see. We love to see it <3. In #149
• Not only did he tackle this bug, @JoeryV found a gap in the contributing docs and updated those to reflect the right state of the world in #155
• @deanna-minnick from dbt labs found a small type when they were testing out the package and fixed the issue themself in #159

PR Action Updates

• Adding JIRA Workflow by @callum-mcdata in #151
• more workflow updates by @callum-mcdata in #156
• More jira fixes by @callum-mcdata in #158

New Contributors

• @JoeryV made their first contribution in #155
• @deanna-minnick made their first contribution in #159

Full Changelog: 1.3.0...1.3.1

Don't Use This Release Unless Working On dbt-core 1.4.0-a1

This release changes the required-dbt-version

1.3.0 - The Coalesce Release

Breaking Changes

• Updating the package to reflect the changes in the v1.3 metric spec.
  • type --> calculation_method
  • sql --> expression
  • type: expression --> calculation_method: derived
• Updating the required_dbt_version to be v1.3
• Removing the dependency on dbt_utils.

New Features

• Adding functionality to return a single value for a metric by using all_time in the time_grains by @callum-mcdata in #99
• Adding a new metric_list input to secondary calculations to allow for flexibility in which metric the secondary calculation is being applied to. By @callum-mcdata in #116
• Making develop have feature parity with calculate! You can now query any metrics defined in the provided yml, even if they are derived metrics. By @callum-mcdata in #118
• Allow users to override the default method for handling null values. Default behavior replaces null values with 0 - now there is a config to keep as null. Shout out to @dave-connors-3 taking on a bear of a PR . #96
• Adding the ability to define a window in the metric definition and have that be applied in the calculation. More information in the readme. By @callum-mcdata in #104
• Adding a new secondary calculation called prior that allows you to display the metric prior metric value from a set interval. By @callum-mcdata in #127

Quality of Life Changes

• The 👑 @dave-connors-3 cleaning up the whitespace in the compiled query so it can actually be read by humans! #94
• Restructuring the metrics in the package for code consistency and to remove branching paths based on different behaviors/inputs. #90
• Adding default ordering to the final returned dataset where the most recent metric value is first! #119
• Updating packages.yml by @callum-mcdata in #122
• Adding changie for our change management by @callum-mcdata in #137

Bug Fixes

• Fixing the calendar table by @callum-mcdata in #98
• Hotfix: All_time with dimensions broken by @callum-mcdata in #108
• Fixing Bug Around Left Outer Join Vs Full Outer Join by @callum-mcdata in https://github.com/dbt-labs/dbt_metrics/pull/133s
• Removing dbt_utils date trunc functionality by @callum-mcdata in #134
• Removing order by for redshift by @callum-mcdata in #136
• Removing L Character in Calculate by @callum-mcdata in #126
• Fixing Rolling Secondary Calc Typo by @callum-mcdata in #141
• Removing bool_or from dbt_utils reference by @callum-mcdata in #142

Full Changelog: 0.3.2...1.3.0