fix: datatype tracking issue on virtual dataset

[codemaster08240328]

SUMMARY

Column types not detected when creating Virtual Dataset.

If we create a virtual dataset, and try to create a chart based on it, some column types may not be detected as it was defined in original dataset. It was because some columns had no any non-null values when we run the SQL to create a dataset.
The previous version was tracking column type of columns from the query result values, and that's why some columns had null type even though it had exact type in original dataset.

To fix this, we need to track the column type from original dataset, not a SQL query running result value.

BEFORE/AFTER SCREENSHOTS OR ANIMATED GIF

BEFORE

AFTER:

TESTING INSTRUCTIONS

1. Create a virtual dataset using this query:

SELECT *
FROM public."SQLLab_Test_For_VirtualDS"
WHERE site_slug in ('connectmiles', 'rocketmiles')
AND reward_program_slug in ('connectmiles')
LIMIT 10

2. Try to create a chart based on virtual dataset.
3. See if all columns have exact data type, not NULL.

Dataset for virtual dataset.
sqllab_query_publicconnectmiles_all_reservation_stats_20220412T210849.csv

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

[codecov]

Codecov Report

Merging #20088 (fa4f374) into master (c8fe518) will decrease coverage by 11.91%.
The diff coverage is 37.50%.

@@             Coverage Diff             @@
##           master   #20088       +/-   ##
===========================================
- Coverage   66.47%   54.55%   -11.92%     
===========================================
  Files        1727     1727               
  Lines       64724    64732        +8     
  Branches     6822     6822               
===========================================
- Hits        43024    35314     -7710     
- Misses      19969    27687     +7718     
  Partials     1731     1731               

Flag	Coverage Δ
hive	53.69% <37.50%> (-0.01%)	⬇️
mysql	?
postgres	?
presto	53.55% <37.50%> (-0.01%)	⬇️
python	57.98% <37.50%> (-24.66%)	⬇️
sqlite	?
unit	49.45% <37.50%> (-0.01%)	⬇️

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

Impacted Files	Coverage Δ
superset/db_engine_specs/postgres.py	59.32% <37.50%> (-37.96%)	⬇️
superset/utils/dashboard_import_export.py	0.00% <0.00%> (-100.00%)	⬇️
superset/key_value/commands/upsert.py	0.00% <0.00%> (-89.59%)	⬇️
superset/key_value/commands/update.py	0.00% <0.00%> (-89.37%)	⬇️
superset/key_value/commands/delete.py	0.00% <0.00%> (-85.30%)	⬇️
superset/key_value/commands/delete_expired.py	0.00% <0.00%> (-80.77%)	⬇️
superset/dashboards/commands/importers/v0.py	15.62% <0.00%> (-76.25%)	⬇️
superset/datasets/commands/update.py	25.88% <0.00%> (-68.24%)	⬇️
superset/datasets/commands/create.py	30.18% <0.00%> (-67.93%)	⬇️
superset/datasets/commands/importers/v0.py	24.03% <0.00%> (-67.45%)	⬇️
... and 272 more

---

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 c8fe518...fa4f374. Read the comment docs.

[betodealmeida]

Looks good! :)

[betodealmeida]

Looks like there's a test failing.

[villebro]

I think this import should be moved into get_datatype, as psycopg2 isn't a required dependency

[michael-s-molina]

Fixed by #20543