[Jinja] Make Presto template functions backwards compatible

[etr2460]

CATEGORY

Choose one

• Bug Fix
• Enhancement (new features, refinement)
• Refactor
• Add tests
• Build / Development Environment
• Documentation

SUMMARY

#7782 changed the behavior of the presto.latest_partition Jinja template function to return an array instead of a string, breaking all sql metrics/slices that used it. This PR makes the latest_partition function behave the same as prior to this change, introduces a new function presto.latest_partitions that can be used with this new behavior, and provides documentation for the change in code.

It might be worth renaming these functions if we were ever to do a breaking release of superset, but i don't see that happening any time soon.

TEST PLAN

Tested locally, CI

ADDITIONAL INFORMATION

• Has associated issue:
• Changes UI
• Requires DB Migration.
• Confirm DB Migration upgrade and downgrade tested.
• Introduces new feature or API
• Removes existing feature or API

REVIEWERS

@michellethomas @john-bodley @betodealmeida @DiggidyDave @xtinec

[john-bodley]

Do we need this function as it's only called once from above. Additionally the logic looks similar to latest_partitions and thus it seems line #253 could simply be,

return self.latest_partitions(table_name)[0]

Note for both latest_partition and latest_partitions an IndexError could be thrown and thus it's probably prudent to add mention in the docstring, i.e.,

:param table_name: table name in the format `schema.table`
:return: the first (or only) value in the latest partition array
:raises IndexError: If no partition exists

[etr2460]

updated

[codecov-io]

Codecov Report

Merging #7993 into master will increase coverage by 0.02%.
The diff coverage is 75%.

@@            Coverage Diff            @@
##           master   #7993      +/-   ##
=========================================
+ Coverage   65.57%   65.6%   +0.02%     
=========================================
  Files         469     469              
  Lines       22414   22417       +3     
  Branches     2431    2431              
=========================================
+ Hits        14699   14707       +8     
+ Misses       7594    7589       -5     
  Partials      121     121

Impacted Files	Coverage Δ
superset/jinja_context.py	76.69% <75%> (-0.31%)	⬇️
superset/utils/core.py	87.83% <0%> (+0.16%)	⬆️
superset/db_engine_specs/sqlite.py	66.66% <0%> (+20.83%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e455f1d...029ef2a. Read the comment docs.

[villebro]

Small change request: There's a reference to latest_partition('some_table') in docs/sqllab.rst; I'm assuming that method is deprecated in favour of first_latest_partition('first table')? I'm hoping to break out the template part into a page of it's own soon as templating isn't really a subset of SQLLab, but until then it would be good to keep the existing docs as up to date as possible.

[etr2460]

@villebro comment addressed!

[villebro]

Hehe thanks @etr2460 sorry for being a pain! 😁

[etr2460]

Not a pain at all! it's important to keep documentation up to date :D