Skip to content

Add custom_metadata to IPC DictionaryEncoding #11622

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.

Sign up for GitHub

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

Already on GitHub? Sign in to your account

Draft
rustyconover wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

Add custom_metadata to IPC DictionaryEncoding #11622

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
274 changes: 216 additions & 58 deletions arrow-integration-test/src/field.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,19 @@ use arrow::error::{ArrowError, Result};
use std::collections::HashMap;
use std::sync::Arc;

/// The metadata key of an extension type name
const EXTENSION_TYPE_NAME_KEY: &str = "ARROW:extension:name";

/// Parse a `Field` definition from a JSON representation.
///
/// The optional `metadata` of a field's `dictionary` holds the metadata of the
/// dictionary value type, for example its extension type. [`DataType::Dictionary`]
/// cannot carry metadata for its value type (see
/// <https://github.com/apache/arrow-rs/issues/7982>), so it is mapped to the
/// metadata of the returned [`Field`] in the same way as the arrow-ipc readers
/// map `DictionaryEncoding.custom_metadata`: it is copied into the field's
/// metadata, without replacing keys already present there, unless the field
/// has its own `ARROW:extension:name`, in which case it is dropped.
pub fn field_from_json(json: &serde_json::Value) -> Result<Field> {
use serde_json::Value;
match *json {
Expand Down Expand Up @@ -49,64 +61,9 @@ pub fn field_from_json(json: &serde_json::Value) -> Result<Field> {
};

// Referenced example file: testing/data/arrow-ipc-stream/integration/1.0.0-littleendian/generated_custom_metadata.json.gz
let metadata = match map.get("metadata") {
Some(Value::Array(values)) => {
let mut res: HashMap<String, String> = HashMap::default();
for value in values {
match value.as_object() {
Some(map) => {
if map.len() != 2 {
return Err(ArrowError::ParseError(
"Field 'metadata' must have exact two entries for each key-value map".to_string(),
));
}
if let (Some(k), Some(v)) = (map.get("key"), map.get("value")) {
if let (Some(k_str), Some(v_str)) = (k.as_str(), v.as_str()) {
res.insert(
k_str.to_string().clone(),
v_str.to_string().clone(),
);
} else {
return Err(ArrowError::ParseError(
"Field 'metadata' must have map value of string type"
.to_string(),
));
}
} else {
return Err(ArrowError::ParseError("Field 'metadata' lacks map keys named \"key\" or \"value\"".to_string()));
}
}
_ => {
return Err(ArrowError::ParseError(
"Field 'metadata' contains non-object key-value pair"
.to_string(),
));
}
}
}
res
}
// We also support map format, because Schema's metadata supports this.
// See https://github.com/apache/arrow/pull/5907
Some(Value::Object(values)) => {
let mut res: HashMap<String, String> = HashMap::default();
for (k, v) in values {
if let Some(str_value) = v.as_str() {
res.insert(k.clone(), str_value.to_string().clone());
} else {
return Err(ArrowError::ParseError(format!(
"Field 'metadata' contains non-string value for key {k}"
)));
}
}
res
}
Some(_) => {
return Err(ArrowError::ParseError(
"Field `metadata` is not json array".to_string(),
));
}
_ => HashMap::default(),
let mut metadata = match map.get("metadata") {
Some(metadata) => metadata_from_json(metadata, "metadata")?,
None => HashMap::default(),
};

// if data_type is a struct or list, get its children
Expand Down Expand Up @@ -278,6 +235,16 @@ pub fn field_from_json(json: &serde_json::Value) -> Result<Field> {
));
}
};
// metadata for the dictionary value type, see above
if let Some(value_metadata) = dictionary.get("metadata") {
let value_metadata =
metadata_from_json(value_metadata, "dictionary.metadata")?;
if !metadata.contains_key(EXTENSION_TYPE_NAME_KEY) {
for (k, v) in value_metadata {
metadata.entry(k).or_insert(v);
}
}
}
DataType::Dictionary(Box::new(index_type), Box::new(data_type))
}
_ => data_type,
Expand All @@ -294,6 +261,69 @@ pub fn field_from_json(json: &serde_json::Value) -> Result<Field> {
}
}

/// Parse the key-value metadata in the field attribute `attribute`.
///
/// The metadata is a list of `{"key": ..., "value": ...}` objects. A map from
/// keys to values is also accepted, because Schema's metadata supports this.
/// See <https://github.com/apache/arrow/pull/5907>
fn metadata_from_json(
json: &serde_json::Value,
attribute: &str,
) -> Result<HashMap<String, String>> {
use serde_json::Value;
match json {
Value::Array(values) => {
let mut res: HashMap<String, String> = HashMap::default();
for value in values {
match value.as_object() {
Some(map) => {
if map.len() != 2 {
return Err(ArrowError::ParseError(format!(
"Field '{attribute}' must have exact two entries for each key-value map"
)));
}
if let (Some(k), Some(v)) = (map.get("key"), map.get("value")) {
if let (Some(k_str), Some(v_str)) = (k.as_str(), v.as_str()) {
res.insert(k_str.to_string(), v_str.to_string());
} else {
return Err(ArrowError::ParseError(format!(
"Field '{attribute}' must have map value of string type"
)));
}
} else {
return Err(ArrowError::ParseError(format!(
"Field '{attribute}' lacks map keys named \"key\" or \"value\""
)));
}
}
_ => {
return Err(ArrowError::ParseError(format!(
"Field '{attribute}' contains non-object key-value pair"
)));
}
}
}
Ok(res)
}
Value::Object(values) => {
let mut res: HashMap<String, String> = HashMap::default();
for (k, v) in values {
if let Some(str_value) = v.as_str() {
res.insert(k.clone(), str_value.to_string());
} else {
return Err(ArrowError::ParseError(format!(
"Field '{attribute}' contains non-string value for key {k}"
)));
}
}
Ok(res)
}
_ => Err(ArrowError::ParseError(format!(
"Field `{attribute}` is not json array"
))),
}
}

/// Generate a JSON representation of the `Field`.
pub fn field_to_json(field: &Field) -> serde_json::Value {
let children: Vec<serde_json::Value> = match field.data_type() {
Expand Down Expand Up @@ -616,4 +646,132 @@ mod tests {

assert_eq!(expected, dt);
}

/// JSON of a dictionary-encoded `fixedsizebinary(16)` field, with the given
/// `dictionary.metadata` and field `metadata` attributes (if any)
fn uuid_dictionary_field_json(
dictionary_metadata: Option<&str>,
metadata: Option<&str>,
) -> Value {
let mut json = serde_json::json!({
"name": "uuids",
"nullable": true,
"type": {"name": "fixedsizebinary", "byteWidth": 16},
"children": [],
"dictionary": {
"id": 0,
"indexType": {"name": "int", "isSigned": true, "bitWidth": 32},
"isOrdered": false
}
});
if let Some(dictionary_metadata) = dictionary_metadata {
json["dictionary"]["metadata"] = serde_json::from_str(dictionary_metadata).unwrap();
}
if let Some(metadata) = metadata {
json["metadata"] = serde_json::from_str(metadata).unwrap();
}
json
}

fn uuid_dictionary_type() -> DataType {
DataType::Dictionary(
Box::new(DataType::Int32),
Box::new(DataType::FixedSizeBinary(16)),
)
}

/// Read a dictionary field with the given `dictionary.metadata` and field
/// `metadata` (as `[{"key": ..., "value": ...}]` lists)
fn read_uuid_dictionary_field(
dictionary_metadata: Option<&[(&str, &str)]>,
metadata: Option<&[(&str, &str)]>,
) -> Field {
let to_json = |pairs: &[(&str, &str)]| {
let pairs = pairs
.iter()
.map(|(k, v)| serde_json::json!({"key": k, "value": v}))
.collect::<Vec<_>>();
serde_json::to_string(&pairs).unwrap()
};
let json = uuid_dictionary_field_json(
dictionary_metadata.map(to_json).as_deref(),
metadata.map(to_json).as_deref(),
);
let field = field_from_json(&json).unwrap();
assert_eq!(field.data_type(), &uuid_dictionary_type());
field
}

#[test]
fn parse_dictionary_value_metadata_from_json() {
let name = EXTENSION_TYPE_NAME_KEY;
let ext_metadata = "ARROW:extension:metadata";
let expected = |pairs: &[(&str, &str)]| {
pairs
.iter()
.map(|(k, v)| (k.to_string(), v.to_string()))
.collect::<HashMap<_, _>>()
};

// copied to the field metadata
let field = read_uuid_dictionary_field(Some(&[(name, "arrow.uuid")]), None);
assert_eq!(field.metadata(), &expected(&[(name, "arrow.uuid")]));

// without replacing the keys of the field metadata
let field = read_uuid_dictionary_field(
Some(&[
(name, "arrow.uuid"),
(ext_metadata, "value"),
("k", "value"),
]),
Some(&[(ext_metadata, "field"), ("k", "field")]),
);
assert_eq!(
field.metadata(),
&expected(&[
(name, "arrow.uuid"),
(ext_metadata, "field"),
("k", "field")
])
);

// dropped if the field has its own extension type
let field = read_uuid_dictionary_field(
Some(&[(name, "arrow.uuid"), ("k", "value")]),
Some(&[(name, "example.uuid_dictionary")]),
);
assert_eq!(
field.metadata(),
&expected(&[(name, "example.uuid_dictionary")])
);

// empty
let field = read_uuid_dictionary_field(Some(&[]), Some(&[("k", "field")]));
assert_eq!(
field,
read_uuid_dictionary_field(None, Some(&[("k", "field")]))
);
}

#[test]
fn parse_invalid_dictionary_value_metadata_from_json() {
for (metadata, error) in [
(
r#""arrow.uuid""#,
"Field `dictionary.metadata` is not json array",
),
(
r#"[{"key": "ARROW:extension:name"}]"#,
"Field 'dictionary.metadata' must have exact two entries for each key-value map",
),
(
r#"[{"key": "ARROW:extension:name", "value": 1}]"#,
"Field 'dictionary.metadata' must have map value of string type",
),
] {
let value = uuid_dictionary_field_json(Some(metadata), None);
let err = field_from_json(&value).unwrap_err();
assert_eq!(err.to_string(), format!("Parser error: {error}"));
}
}
}
4 changes: 4 additions & 0 deletions arrow-integration-test/src/lib.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -149,6 +149,10 @@ impl From<&Field> for ArrowJsonField {
}

/// Represents a dictionary-encoded field in the Arrow JSON format
///
/// The optional `metadata` of the dictionary value type is not kept, so it is
/// not seen when converting an [`ArrowJsonField`] to a [`Field`]. Use
/// [`field_from_json`] to read it.
#[derive(Deserialize, Serialize, Debug)]
pub struct ArrowJsonFieldDictionary {
/// A unique identifier for the dictionary
Expand Down
Loading