Commit adds a WHERE clause to the sql statement that allows for
arbitrary batching in a given table.

When multiple table checks are given to the SQLTableCheckOperator
and at least one is not a fully aggregate statement, a GROUP BY
clause was previously needed. This commit updates the operator to
use the get_pandas_df() method instead of _get_first() to return a
pandas dataframe object that contains the check names and check
results from the new style of query. The new style of query uses
UNION ALL to run each test as its own SELECT statement, bypassing
the need to do a GROUP BY.

Changed name of method from _get_failed_tests to _get_failed_checks
to better match naming, and updated logic of the method to include
an optional column param. The query in the column check operator
is removed from the failed test exception message, as it was only
ever showing the last query, instead of the relevant one(s). This is
replaced by the column, which will be more useful in debugging.

Without a table alias, the query does not run on Postgres and
other databases. The alias is arbitrary and used only for
proper query execution.

Commit adds a WHERE clause to the sql statement that allows for
arbitrary batching in a given table.

When multiple table checks are given to the SQLTableCheckOperator
and at least one is not a fully aggregate statement, a GROUP BY
clause was previously needed. This commit updates the operator to
use the get_pandas_df() method instead of _get_first() to return a
pandas dataframe object that contains the check names and check
results from the new style of query. The new style of query uses
UNION ALL to run each test as its own SELECT statement, bypassing
the need to do a GROUP BY.

Changed name of method from _get_failed_tests to _get_failed_checks
to better match naming, and updated logic of the method to include
an optional column param. The query in the column check operator
is removed from the failed test exception message, as it was only
ever showing the last query, instead of the relevant one(s). This is
replaced by the column, which will be more useful in debugging.

Without a table alias, the query does not run on Postgres and
other databases. The alias is arbitrary and used only for
proper query execution.

…thub.com:denimalpaca/airflow into sql_check_operators_various_functionality_update

The table alias should be in the self.sql query build statement
as that is where the table it needs to alias is defined.

Commit adds a WHERE clause to the sql statement that allows for
arbitrary batching in a given table.

When multiple table checks are given to the SQLTableCheckOperator
and at least one is not a fully aggregate statement, a GROUP BY
clause was previously needed. This commit updates the operator to
use the get_pandas_df() method instead of _get_first() to return a
pandas dataframe object that contains the check names and check
results from the new style of query. The new style of query uses
UNION ALL to run each test as its own SELECT statement, bypassing
the need to do a GROUP BY.

Changed name of method from _get_failed_tests to _get_failed_checks
to better match naming, and updated logic of the method to include
an optional column param. The query in the column check operator
is removed from the failed test exception message, as it was only
ever showing the last query, instead of the relevant one(s). This is
replaced by the column, which will be more useful in debugging.

Without a table alias, the query does not run on Postgres and
other databases. The alias is arbitrary and used only for
proper query execution.

…thub.com:denimalpaca/airflow into sql_check_operators_various_functionality_update

Fixed bug in test where the dataframe column names did not match
the operator's expected dataframe column names. Added more info
to the SQLColumnCheckOperator's batch arg. Fixed the location of
table aliasing in SQLTableCheckOperator.

Gives a clearer name to the parameter and adds templating to
the SQLTableCheckOperator.