[Bug Fix] Removing escape_sql so we dont double escape

[michellethomas]

@john-bodley Removing escape_sql because this seems to be handled by qry.compile(engine) here and both of them escaping double escapes.

[michellethomas]

@mistercrunch we are a little unsure what would have caused this regression, as nothing seems to have changed in the related code recently. It looks like in sqlalchemy 1.2 they did some work to their autoescape feature. Maybe our sqlalchemy upgrade caused this issue, but it's a little unclear. Any thoughts on why we might be seeing double % escaping? Anything else we should be aware of in removing this part of the code?

[codecov-io]

Codecov Report

Merging #4590 into master will increase coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #4590      +/-   ##
==========================================
+ Coverage   71.17%   71.18%   +<.01%     
==========================================
  Files         188      188              
  Lines       14834    14828       -6     
  Branches     1086     1086              
==========================================
- Hits        10558    10555       -3     
+ Misses       4273     4270       -3     
  Partials        3        3

Impacted Files	Coverage Δ
superset/db_engine_specs.py	52.26% <ø> (+0.01%)	⬆️
superset/connectors/sqla/models.py	77.91% <ø> (+0.11%)	⬆️

---

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 f988110...e1af421. Read the comment docs.

[john-bodley]

@michellethomas upon further investigation this regression doesn't seem to be related to the auto-escape feature but rather this SQLAlchemy commit (released in version 1.2.0) which escapes percent signs (%) for all dialects where the param style is either format or pyformat per this statement which is true for Presto. Hence by upgrading SQLAlchemy from 1.1.9 to 1.2.2 (PR) the percent signs were doubly escaped in Presto per the PrestoEngineSpec causing the regression.