chore: add check constraint to restrict `Slice` models datasource_type != "table" #23614

hughhhh · 2023-04-06T15:29:57Z

SUMMARY

Currently experiencing errors in production where charts/slices are being saved with datasource_type = "query". Currently charts can only be powered by SqlaTable models.

BEFORE/AFTER SCREENSHOTS OR ANIMATED GIF

TESTING INSTRUCTIONS

ADDITIONAL INFORMATION

Has associated issue:
Required feature flags:
Changes UI
Includes DB Migration (follow approval process in SIP-59)
- Migration is atomic, supports rollback & is backwards-compatible
- Confirm DB migration upgrade and downgrade tested
- Runtime estimates and downtime expectations provided
Introduces new feature or API
Removes existing feature or API

…t-ds-constraint

john-bodley · 2023-04-06T18:27:50Z

@hughhhh at Airbnb we're seeing examples of this as well. My question is your fix only includes a migration, i.e., it will remedy any currently ill-defined charts, but there's no fix for actually preventing this again—beyond the check constraint? Has this already been resolved?

codecov · 2023-04-06T18:54:49Z

Codecov Report

Merging #23614 (7ac7868) into master (bccd267) will increase coverage by 0.36%.
The diff coverage is 74.22%.

❗ Current head 7ac7868 differs from pull request most recent head 6aee98b. Consider uploading reports for the commit 6aee98b to get more accurate results

@@            Coverage Diff             @@
##           master   #23614      +/-   ##
==========================================
+ Coverage   67.64%   68.00%   +0.36%     
==========================================
  Files        1916     1920       +4     
  Lines       74036    73990      -46     
  Branches     8039     8091      +52     
==========================================
+ Hits        50078    50314     +236     
+ Misses      21911    21607     -304     
- Partials     2047     2069      +22

Flag	Coverage Δ
hive	`53.18% <50.70%> (?)`
mysql	`79.21% <74.82%> (+0.72%)`	⬆️
postgres	`?`
presto	`53.09% <50.70%> (+0.41%)`	⬆️
python	`82.98% <75.53%> (+0.74%)`	⬆️
sqlite	`?`
unit	`53.03% <42.90%> (+0.38%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...d/packages/superset-ui-chart-controls/src/types.ts	`100.00% <ø> (ø)`
...ackages/superset-ui-core/src/utils/featureFlags.ts	`100.00% <ø> (ø)`
...s/plugin-chart-echarts/src/Timeseries/constants.ts	`100.00% <ø> (ø)`
...gin-chart-echarts/src/Timeseries/transformProps.ts	`57.27% <0.00%> (-1.07%)`	⬇️
...tend/plugins/plugin-chart-echarts/src/constants.ts	`100.00% <ø> (ø)`
...tend/plugins/plugin-chart-echarts/src/controls.tsx	`76.00% <ø> (ø)`
...ugins/preset-chart-xy/src/components/Line/Line.tsx	`0.00% <0.00%> (ø)`
...preset-chart-xy/src/utils/createMarginSelector.tsx	`0.00% <0.00%> (ø)`
superset-frontend/src/SqlLab/App.jsx	`0.00% <0.00%> (ø)`
...d/src/SqlLab/components/SaveDatasetModal/index.tsx	`52.27% <0.00%> (ø)`
... and 64 more

... and 139 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

john-bodley · 2023-04-06T18:28:34Z

superset/migrations/versions/2023-03-27_12-30_7e67aecbf3f1_chart_ds_constraint.py

+    __tablename__ = "slices"
+
+    id = sa.Column(sa.Integer, primary_key=True)
+    slice_name = sa.Column(sa.String(250))


This is unused.

john-bodley · 2023-04-06T18:28:41Z

superset/migrations/versions/2023-03-27_12-30_7e67aecbf3f1_chart_ds_constraint.py

+    id = sa.Column(sa.Integer, primary_key=True)
+    slice_name = sa.Column(sa.String(250))
+    datasource_type = sa.Column(sa.String(200))
+    query_context = sa.Column(sa.Text)


This is unused.

john-bodley · 2023-04-06T18:51:42Z

superset/migrations/versions/2023-03-27_12-30_7e67aecbf3f1_chart_ds_constraint.py

+    with op.batch_alter_table("slices") as batch_op:
+        for slc in session.query(Slice).filter(Slice.datasource_type != "table").all():
+            # clean up all charts with datasource_type not != table
+            slc.datasource_type = "table"


You'll also have to update the params which has datasource field defined as <datasource_id>_<datasource_type>.

@hughhhh we're seeing a defined slices.datasource_name column when slices.datasource_type = 'query' however it doesn't seem to be a dataset in the tables table—which would indicate that your migration would likely break said slices when you make the switch from query to table. BTW where is this query stored?

@hughhhh we're seeing a defined slices.datasource_name column when slices.datasource_type = 'query' however it doesn't seem to be a dataset in the tables table—which would indicate that your migration would likely break said slices when you make the switch from query to table. BTW where is this query stored?

So those slices are most likely broken, because charts/slices don't know who to execute grabbing data when the datasource_type is Query

Query is defined here

this is the stack trace we are encountering in production:

return [ File "/usr/local/lib/python3.8/site-packages/marshmallow/schema.py", line 515, in <listcomp> self._serialize(d, many=False) File "/usr/local/lib/python3.8/site-packages/marshmallow/schema.py", line 520, in _serialize value = field_obj.serialize(attr_name, obj, accessor=self.get_attribute) File "/usr/local/lib/python3.8/site-packages/marshmallow/fields.py", line 338, in serialize return self._serialize(value, attr, obj, **kwargs) File "/usr/local/lib/python3.8/site-packages/marshmallow/fields.py", line 1922, in _serialize return self._call_or_raise(self.serialize_func, obj, attr) File "/usr/local/lib/python3.8/site-packages/marshmallow/fields.py", line 1936, in _call_or_raise return func(value) File "/usr/local/lib/python3.8/site-packages/superset/models/slice.py", line 167, in datasource_url return datasource.explore_url if datasource else None AttributeError: 'Query' object has no attribute 'explore_url'

@hughhhh per your response,

So those slices are most likely broken, because charts/slices don't know who to execute grabbing data when the datasource_type is Query

I guess the question/concern is simply changing the datasource_type to table doesn't really mitigate the problematic charts as the datasource_id is then likely incorrect as well, i.e., it's referencing the query.id column rather than the tables.id column.

Also please refer to my comment regarding the chart parameters as we (for right or wrong) doubly define the ID/type tuple.

so the idea here was to try and see if some of them will work if we change the type, since we can't just delete these charts. We are assuming these charts are failing already and hoping this might help fix it, but no guarantee. Also now with the constraint we should see error when the user/api tries to write slice.datasource_type = query

…e != "table" (apache#23614)

hughhhh added 3 commits March 27, 2023 12:31

add migration

564ae4a

lit

4d673a6

Merge branch 'master' of https://github.com/apache/superset into char…

237e2ec

…t-ds-constraint

hughhhh requested a review from a team as a code owner April 6, 2023 15:29

pull-request-size bot added the size/M label Apr 6, 2023

update table name

a84bde9

hughhhh force-pushed the chart-ds-constraint branch from 5afaf6e to a84bde9 Compare April 6, 2023 15:42

update

d552b42

john-bodley reviewed Apr 6, 2023

View reviewed changes

hughhhh added 2 commits April 10, 2023 12:45

update filter

7baf330

added tests

6aee98b

pull-request-size bot added size/L and removed size/M labels Apr 13, 2023

john-bodley approved these changes Apr 20, 2023

View reviewed changes

hughhhh merged commit c441a70 into master Apr 20, 2023

rusackas deleted the chart-ds-constraint branch April 20, 2023 21:24

sebastianliebscher pushed a commit to sebastianliebscher/superset that referenced this pull request Apr 28, 2023

chore: add check constraint to restrict Slice models datasource_typ…

adb2370

…e != "table" (apache#23614)

mistercrunch added 🏷️ bot A label used by `supersetbot` to keep track of which PR where auto-tagged with release labels 🚢 3.0.0 labels Mar 13, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

chore: add check constraint to restrict `Slice` models datasource_type != "table" #23614

chore: add check constraint to restrict `Slice` models datasource_type != "table" #23614

hughhhh commented Apr 6, 2023

john-bodley commented Apr 6, 2023 •

edited

Loading

codecov bot commented Apr 6, 2023 •

edited

Loading

john-bodley Apr 6, 2023

john-bodley Apr 6, 2023

john-bodley Apr 6, 2023

john-bodley Apr 6, 2023

hughhhh Apr 6, 2023 •

edited

Loading

hughhhh Apr 6, 2023

john-bodley Apr 10, 2023

hughhhh Apr 13, 2023

chore: add check constraint to restrict Slice models datasource_type != "table" #23614

chore: add check constraint to restrict Slice models datasource_type != "table" #23614

Conversation

hughhhh commented Apr 6, 2023

SUMMARY

BEFORE/AFTER SCREENSHOTS OR ANIMATED GIF

TESTING INSTRUCTIONS

ADDITIONAL INFORMATION

john-bodley commented Apr 6, 2023 • edited Loading

codecov bot commented Apr 6, 2023 • edited Loading

Codecov Report

john-bodley Apr 6, 2023

Choose a reason for hiding this comment

john-bodley Apr 6, 2023

Choose a reason for hiding this comment

john-bodley Apr 6, 2023

Choose a reason for hiding this comment

john-bodley Apr 6, 2023

Choose a reason for hiding this comment

hughhhh Apr 6, 2023 • edited Loading

Choose a reason for hiding this comment

hughhhh Apr 6, 2023

Choose a reason for hiding this comment

john-bodley Apr 10, 2023

Choose a reason for hiding this comment

hughhhh Apr 13, 2023

Choose a reason for hiding this comment

chore: add check constraint to restrict `Slice` models datasource_type != "table" #23614

chore: add check constraint to restrict `Slice` models datasource_type != "table" #23614

john-bodley commented Apr 6, 2023 •

edited

Loading

codecov bot commented Apr 6, 2023 •

edited

Loading

hughhhh Apr 6, 2023 •

edited

Loading