✨ New ✨
Mode Integration through JDBC
Details
Product Component
Interfaces - SQL
More Detail
We've released the production version of our integration with Mode analytics. The integration will enable Mode users to have the flexibility of reporting and BI of Mode, with the assurance of accurate metrics from Transform.
Our integration allows you to express MQL requests through SQL using a JDBC interface.
Example:
SELECT *
FROM
MQL_QUERY(<
bookings
BY
user__country>);
Verification Steps
To verify this, you need a Mode analytics instance.
Documentation for setup with Mode can be found here .
Documentation for using our JDBC interface can be found here.
Documentation Site
Details
Product Component
General
More Detail
We are excited to announce that we are officially live with a documentation site at docs.transform.co which you can access from our product. Above the setting link on the left navigation, you'll find a new link called "Documentation".
Note some of our documentation is a work in progress. Please contact us if there is specific content you'd like to see!
Verification Steps
Use the new left nav "Documentation" link to access Transform's Documentation Site.
Quarterly Granularity
Details
Product Component
Metrics Catalog
More Detail
We have now released Quarterly granularity as an option in our chart menu, which will aggregate values by the calendar quarter.
Verification Steps
Open up the granularity menu in Edit Chart and confirm that Quarterly is an option!
Surfacing Local Dimensions
Details
Product Component
Metrics Catalog
More Detail
Certain data sources do not have primary identifiers and so those dimensions cannot be joined onto other data sources. However, those dimensions can still be accessed locally to that data source. Previously, our UI hid these dimensions, but now we list these dimensions at the bottom of the menu under "Local Dimensions". You can optionally search "Local" to find all these dimensions
Verification Steps
For a metric that's based on a data source without a primary identifier and with dimensions, check that you can see all the local dimensions at the bottom.
🐛 Bug Fixes 🐛
Cumulative Metric Bug Fix
Details
Product Component
Metrics Framework
More Detail
We fixed a bug where cumulative metrics were returning incorrect results for some situations
The issue was a spurious groupby when Transform renders the time range join for cumulative metrics. Cumulative metrics are implemented by aggregating the metric's measure across the window specified in the metric. For example:
select
sum(bookings) as bookings
from bookings_table a
join bookings_table b
on b.ds <= a.ds and b.ds > date_sub(a.ds, 7, days)
group by b.ds, bookings
In this case, the time range (b.ds) was non-varying, Transform would incorrectly group all the bookings values across those 7 days (removing duplicated values). Grouping by a.ds, bookings would work but would be a no-operative, so the fix removed the group by entirely.
Verification Steps
Create a cumulative metric such that the data being accumulated contains repeated values.
ie "weekly bookings"
metric:
name: weekly_bookings
description: weekly bookings
owners:
- bobby@transformdata.io
type: cumulative
type_params:
measures:
- bookings
window: 7 days
measure_bookings, ds
1000 2020/01/01
2000 2020/01/02
1000 2020/01/03
3000 2020/01/04
Previously calculating weekly_bookings for 2020-01-05 would yield 6000 (omitting the repeated 1k value). It would now yield the correct value (7000)
No longer formatting dimension values
Details
Product Component
Metrics Catalog
More Detail
Previously, we would autoformat your dimension values by removing special characters and adding spaces. This could be confusing because what we displayed in the UI is different than the source database. For any net new Transform deployments, dimension values will not be formatted. Note: Dimension names will still be formatted.
If you would like, you can request to go back to the dimension value auto-formatting behavior. Please reach out to us at support@transformdata.io or over Slack.
Verification Steps
To verify, you can confirm your dimension values in the filter menu match the values in your database.