Implement parsing for ParquetWriterOptions

[alamb]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.
We are implementing configurable parquet writing in DataFusion

We want to be able to allow users to specify the parquet writing options (like compression) via a string like

set parquet.writer_version = 2.0
set parquet.compression = zstd(5)

Describe the solution you'd like
Implement FromStr for the following structures, with some tests.

• Encoding
• Compression
• WriterVersion
• EnabledStatistics

Each of these can be done via a separate PR

The basic code can probably be ported from DataFusion here and add some unit tests: https://github.com/apache/arrow-datafusion/blob/ed85abbb878ef3d60e43797376cb9a40955cd89a/datafusion/core/src/datasource/file_format/parquet.rs#L13

Bonus points for good error messages that give example values (like "Invalid encoding. Valid values: plain_dictionary, rle, etc

Describe alternatives you've considered

Additional context

@devinjdangelo implemented parsing for these in apache/datafusion#7244 however, I think these features could be more generally useful to others

[fansehep]

Looks you want the last like this:

let writer_properties = WriterProperties::from_str("
  set parquet.writer_version = 2.0
  set parquet.compression = zstd(5)").unwrap();

I want to try it. and should we really need the prefix word 'set' and parquet.* ?

[alamb]

I want to try it.

Hi @fansehep -- thank you.

should we really need the prefix word 'set' and parquet.* ?

No. I am sorry for the confusion, the set ... terminology is from DataFusion. This ticket in arrow-rs only covers the parsing of the values

so for example, we would implement FromStr for https://docs.rs/parquet/45.0.0/parquet/format/struct.Encoding.html would allow something like this:

// call `parse` to implement FromStr
let encoding: Encoding = "PLAIN".parse().unwrap();
// call parse again, case insensitive
let encoding: Encoding = "plain".parse().unwrap();

[fansehep]

so for example, we would implement FromStr for https://docs.rs/parquet/45.0.0/parquet/format/struct.Encoding.html would allow something like this:例如，我们为 https://docs.rs/parquet/45.0.0/parquet/format/struct.Encoding.html 实现 FromStr 将允许这样的事情：

// call `parse` to implement FromStr
let encoding: Encoding = "PLAIN".parse().unwrap();
// call parse again, case insensitive
let encoding: Encoding = "plain".parse().unwrap();

Thanks for your help. 😃

[alamb]

Thanks for your help. 😃

Thank YOU!

[raulcd]

This one seems to be finished by:

• feat: support parsing for parquet writer option #4938

@alamb is there anything missing from what you were expecting? Can we close it?

[alamb]

I agree this looks reasonable to me.

Thanks @raulcd and @fansehep