The parser directory contains the files for automated parser tests.
This is organised first into folders for each dialect (e.g. ansi, mysql)
which each then contain both .sql files and .yml files. The intent for these
folders is that each test should be in the highest dialect that it can be in. i.e.
If it can be in the ansi dialect then it should be in there.
Within each folder, any .sql files will be tested that they can
successfully parse (i.e. that they do not raise any errors
and that the parsed result does not contain any unparsable segments).
If there is a .yml file with the same filename as the .sql file
then the structure of the parsed query will also be compared against the structure
within that yaml file.
For best test coverage, add both a .sql and .yml file. The easiest way to
add a .yml file is to run:
python test/generate_parse_fixture_yml.py [--dialect <dialect>] [--filter <glob_filter>] [--new-only]
Or via tox:
tox generate-fixture-yml
Or via tox with arguments:
tox generate-fixture-yml -- --dialect <dialect>
This will regenerate all the parsed structure yml files, or a subset based on the given filters.
To avoid running the whole test suite with tox after changing parsers, you can instead run:
pytest test/dialects/dialects_test.py
to save some time.