Improve Docu and coding environment for tblproperties and features

[datanikkthegreek]

Description

Use Case

Let's say I am a user who wants to set some table properties or features and I have no ideas which feautures exist. Currently it was really hard for me to find this.

My understanding is that we have to options using TableAlterer.addFeature or set set_table_properties

Related Issue(s)

Here some propositions/ questions to trigger improvements :)

For set_table_properties:

• How do I know which properties are allowed/exist. I assume the accepted once are here:

  delta-rs/crates/core/src/table/config.rs

  Line 139 in 8acfa3f

  Self::EnableChangeDataFeed => "delta.enableChangeDataFeed",

• We should document them somewhere and add the link to the API reference
• You could also improve the error message to maybe provide all accepted values?

For addFeature:

• Improve the example. For me this didn't work as you need to import also TableFeautures (missing in the example)
• Shouldn't allow_protocol_version_increase be default True? Why don't we have it for set_table_properties for consistancy?
• Again the questions which options are allows? I had to dive into the code: https://github.com/delta-io/delta-rs/blob/main/python/deltalake/_internal.pyi#L29 In this case at least we can see it by TableFeature.* and the options pop up :) Docu might still help
• It's also hard to know which are currently supported. E.g. Deletion Vectors isn't. Issue it only gives me a bad error when allow_protocol_version_increase=True which is default False. Ok now I was able to set it and second run it gives me an error. Something seems odd here.
• Btw are deletion vectors now supported by deltalake? Last week I saw the issue still open

[ion-elgreco]

• It's also hard to know which are currently supported. E.g. Deletion Vectors isn't. Issue it only gives me a bad error when allow_protocol_version_increase=True which is default False. Ok now I was able to set it and second run it gives me an error. Something seems odd here.

This is normal, the second run fails because we don't support handling those tables.

• Btw are deletion vectors now supported by deltalake? Last week I saw the issue still open

They aren't supported

We are always open to PRS that improves the docs :)

[datanikkthegreek]

@ion-elgreco Hey thanks for the suit reply. :)

I didn't understand your comment about you don't support handling those tables in the second run. How can you support sth the first run and not the second run?

If the deletion vectors are not supported, how am I able to set the deletion vector config? That's really misleading for the users or not?

Let me better understand the module first and then I am happy to support. But I see usability in the backend there might be some improvements which could help. :)

[ion-elgreco]

@ion-elgreco Hey thanks for the suit reply. :)

I didn't understand your comment about you don't support handling those tables in the second run. How can you support sth the first run and not the second run?

Because before the first run it's not active yet, so you can activate it. But the moment you do that, you can't read or write to the table anymore.