Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add example of converting RecordBatches to JSON objects #5364

Merged
merged 4 commits into from
Feb 5, 2024
Merged
Changes from 2 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 29 additions & 1 deletion arrow-json/src/writer.rs
Original file line number Diff line number Diff line change
Expand Up @@ -74,7 +74,35 @@
//! [`LineDelimitedWriter`] and [`ArrayWriter`] will omit writing keys with null values.
//! In order to explicitly write null values for keys, configure a custom [`Writer`] by
//! using a [`WriterBuilder`] to construct a [`Writer`].

//!
//! ## Writing to [serde_json] JSON Objects
//!
//! To serialize [`RecordBatch`]es into an array of
//! [JSON](https://docs.serde.rs/serde_json/) objects, use the [RawValue] api
//!
//! [RawValue]: https://docs.rs/serde_json/latest/serde_json/value/struct.RawValue.html
//!
//! ```
//! # use std::sync::Arc;
//! # use arrow_array::{Int32Array, RecordBatch};
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am not sure how much value this example has, to be honest, other than to demonstrate feature parity with previous releases

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I similarly am not immensely convinced of its utility

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

FWIW I did put this example at the end of the docs, so hopefully it is minimally confusing

//! # use arrow_schema::{DataType, Field, Schema};
//! # use serde_json::{Map, Value};
//!
//! let schema = Schema::new(vec![Field::new("a", DataType::Int32, false)]);
//! let a = Int32Array::from(vec![1, 2, 3]);
//! let batch = RecordBatch::try_new(Arc::new(schema), vec![Arc::new(a)]).unwrap();
//!
//! let json_rows: Vec<Map<String, Value>> = todo!("How do we do this?");
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@tustvold can you help / point me at code that does what you are thinking of so I can update the example?

I couldn't immediately see how to apply the suggestion you are making

Copy link
Contributor

@tustvold tustvold Feb 4, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You can "parse" a serialized JSON string into a RawValue, this allows embedding it into existing serde flows without paying additional decoding overheads. There is no way to obtain a Value, other than to parse the serialized JSON string, this is expected. If this is insufficient for people's use-cases I would suggest they file a ticket with their requirements

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Got it -- I will try and update the example to show reparsing the string to Json value with a note about performance.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Given I am still very confused about how the RawValue api fits in here (perhaps because as you hint, there is no clear usecase), I am going to remove mention from the docs to avoid confusion.

I wonder if people potentially were using the json_serde values as an intermediate representation to map RecordBatches to their own data structures via serde 🤔

Maybe we can point them to the https://crates.io/crates/serde_arrow crate for that usecase 🤔

Copy link
Contributor

@tustvold tustvold Feb 4, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Say you have a larger JSON document you want to embed the arrow data into, you could parse into RawValue in order to embed it. That's the major use-case I can think of

I wonder if people potentially were using the json_serde values as an intermediate representation to map RecordBatches to their own data structures via serde

I guess we shall find out 😅

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I made a PR to serde_arrow with an example of how to use that crate to make arrow arrays out of rust structs: chmp/serde_arrow#131

So now I feel quite good about directing people there ❤️

//! // let json_rows = arrow_json::writer::record_batches_to_json_rows(&[&batch]).unwrap();
//! assert_eq!(
//! serde_json::Value::Object(json_rows[1].clone()),
//! serde_json::json!({"a": 2}),
//! );
//! ```
//!
//!
//!
//!
mod encoder;

use std::iter;
Expand Down
Loading