feat: add optional JSON functions support

Add `datafusion-functions-json` as an optional feature (`json`), giving
Python users `json_extract`, `json_get`, `->`, `->>` and other JSON
operators in SQL queries.

When built with `--features json`, JSON functions are automatically
registered with every SessionContext. Default builds are unaffected.

Changes:
- Add `datafusion-functions-json` to workspace dependencies
- Add optional dependency and `json` feature flag to core crate
- Register JSON functions in SessionContext creation when feature is enabled

Author: crm26

Cargo.toml

@@ -48,6 +48,7 @@ datafusion-common = { version = "53", default-features = false }
 datafusion-functions-aggregate = { version = "53" }
 datafusion-functions-window = { version = "53" }
 datafusion-expr = { version = "53" }
+datafusion-functions-json = { version = "0.53" }
 prost = "0.14.3"
 serde_json = "1"
 uuid = { version = "1.23" }

crates/core/Cargo.toml

@@ -53,6 +53,7 @@ datafusion = { workspace = true, features = ["avro", "unicode_expressions"] }
 datafusion-substrait = { workspace = true, optional = true }
 datafusion-proto = { workspace = true }
 datafusion-ffi = { workspace = true }
+datafusion-functions-json = { workspace = true, optional = true }
 prost = { workspace = true } # keep in line with `datafusion-substrait`
 serde_json = { workspace = true }
 uuid = { workspace = true, features = ["v4"] }
@@ -74,6 +75,7 @@ pyo3-build-config = { workspace = true }
 
 [features]
 default = ["mimalloc"]
+json = ["dep:datafusion-functions-json"]
 protoc = ["datafusion-substrait/protoc"]
 substrait = ["dep:datafusion-substrait"]
 

crates/core/src/context.rs

@@ -391,6 +391,14 @@ impl PySessionContext {
             .with_default_features()
             .build();
         let ctx = Arc::new(SessionContext::new_with_state(session_state));
+
+        // Register JSON functions (json_extract, json_get, ->, ->>) when feature is enabled
+        #[cfg(feature = "json")]
+        datafusion_functions_json::register_all(ctx.as_ref())
+            .map_err(|e| PyErr::new::<pyo3::exceptions::PyRuntimeError, _>(
+                format!("Failed to register JSON functions: {e}"),
+            ))?;
+
         let logical_codec = Self::default_logical_codec(&ctx);
         Ok(PySessionContext { ctx, logical_codec })
     }