Generalize caching of connection in DbApiHook to improve performance

[dabla]

Added a cached connection property in DbApiHook to improve performance (it already existed for MySQLHook and MSSQLHook), also move the duplicated connection_extra_lower property from JdbcHook and OdbcHook to DbApiHook. This enchancement will fix the unnecessary connection lookups for Jdbc and Odbc driver properties.

---

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in newsfragments.

[dabla]

I'm just wondering if I have to bump to version of the common sql provider? As there will be a change there, it should become like 1.15.0 as @potiuk suggested in the other PR 40665. Has this to be done manually (I don't think so but could), I've also read the docs here but didn't find the answer to that.

[dabla]

One test failing due to:

AttributeError: type object 'SkipDBTestsSession' has no attribute 'get_bind'

[dabla]

One test failing due to:

AttributeError: type object 'SkipDBTestsSession' has no attribute 'get_bind'

Seems my other PR are also having the same issue

[eladkal]

Looks like we need now to set min version for apache-airflow-providers-common-sql as 1.17.0
otherwise LGTM

[eladkal]

@dabla can you make the needed change so we can merge this for next release?

[dabla]

@dabla can you make the needed change so we can merge this for next release?

Will do that tomorrow first thing in the morning

[potiuk]

Hey @dabla -> you need to also upgrade common.sql version in their provider.yaml - otherwise PROD image build is not able to find common.sql >= 1.17.0.dev0

The build is smart enough (or it should be at least) to figure out that in this case the common.sql package should be buillt and used for PROD image build - but you need to bump the version in it's provider.yaml.

[dabla]

Hey @dabla -> you need to also upgrade common.sql version in their provider.yaml - otherwise PROD image build is not able to find common.sql >= 1.17.0.dev0

The build is smart enough (or it should be at least) to figure out that in this case the common.sql package should be buillt and used for PROD image build - but you need to bump the version in it's provider.yaml.

Aha, that's why the image build is failing, thanks for letting me know @potiuk, will do that

[dabla]

Hey @dabla -> you need to also upgrade common.sql version in their provider.yaml - otherwise PROD image build is not able to find common.sql >= 1.17.0.dev0

The build is smart enough (or it should be at least) to figure out that in this case the common.sql package should be buillt and used for PROD image build - but you need to bump the version in it's provider.yaml.

@potiuk What about the release notes and stuff for common sql, as when I bump in provider.yaml, do I also need to bump that version init.py file and release notes? Or is there a (breeze) command to bump the version of a provider? Didn't directly find it in docs but that's probably my bad.

[potiuk]

Just add version in the provider.yaml. all the rest should happen by pre-commits

You might want to add a line at top of changelog explaining better what was added in common.sql (but this is usually only done for breaking changes - generally commit messages are added to the changelog when release manager prepares the release.

[dabla]

Just add version in the provider.yaml. all the rest should happen by pre-commits

You might want to add a line at top of changelog explaining better what was added in common.sql (but this is usually only done for breaking changes - generally commit messages are added to the changelog when release manager prepares the release.

Nice, ok thank you @potiuk , wanted to make sure before bumping the version, will add a comment line in top of file explaining what was changed, always better to document anyway and good for the changelog history.

[potiuk]

Nice, ok thank you @potiuk , wanted to make sure before bumping the version, will add a comment line in top of file explaining what was changed, always better to document anyway and good for the changelog history.

Not sure if it will work with out release management preparation automation if you added the Features header (normally it is added automatically when provider bulk documentation is being prepared, but we might see if it works. CC: @eladkal @amoghrajesh

[dabla]

Nice, ok thank you @potiuk , wanted to make sure before bumping the version, will add a comment line in top of file explaining what was changed, always better to document anyway and good for the changelog history.

Not sure if it will work with out release management preparation automation if you added the Features header (normally it is added automatically when provider bulk documentation is being prepared, but we might see if it works. CC: @eladkal @amoghrajesh

If it causes issues, I can revert it as I added this in a dedicated commit.

[potiuk]

If it causes issues, I can revert it as I added this in a dedicated commit.

We will see and RM can modify it on-line. How it works is that it simply updates the changelog locally and RM commits it, so if it causes issues, it can be manually modified before running the script and classifying the issues. Or we can implement a fix to allow that in the future.