[Analytics Backend / DataFusion] Onboard PPL parse to DataFusion

sandbox/libs/analytics-framework/src/main/java/org/opensearch/analytics/spi/FieldType.java

@@ -56,6 +56,8 @@ public enum FieldType {
     OBJECT("object"),
     FLAT_OBJECT("flat_object"),
     COMPLETION("completion"),
+
+    // ── Composite ────────────────────────────────────────────────────
     /**
      * Array-typed expression result. Used for the return-type slot of array-producing scalar
      * functions (PPL {@code array(…)}, {@code array_slice}, {@code array_distinct}). Has no
@@ -67,15 +69,16 @@ public enum FieldType {
     ARRAY("array"),
 
     /**
-     * Map-typed expression result. First in-tree producer is PPL `spath`'s auto-extract mode
-     * (`JSON_EXTRACT_ALL` returns {@code MAP<VARCHAR, VARCHAR>}). Mapping string is {@code
-     * "map"} as a placeholder — no OpenSearch storage format declares this mapping today, so
-     * {@link #fromMappingType} never resolves to it through the mapping path; columns reach
-     * MAP only through {@link #fromSqlTypeName}. Capability registrations for filter / project
-     * operators on MAP columns are intentionally minimal: callers (e.g. PPL `where doc.user.name`)
-     * always wrap the MAP column in an ITEM lookup whose result type is the map's value type,
-     * so the EQUALS / sort / aggregate operators see the value-level type by the time the
-     * runtime executes them.
+     * Map-typed expression result. In-tree producers include PPL {@code spath}'s auto-extract
+     * mode ({@code JSON_EXTRACT_ALL} returns {@code MAP<VARCHAR, VARCHAR>}) and PPL
+     * {@code parse} (named regex groups returned as {@code MAP<VARCHAR, VARCHAR>}). Mapping
+     * string is {@code "map"} as a placeholder — no OpenSearch storage format declares this
+     * mapping today, so {@link #fromMappingType} never resolves to it through the mapping
+     * path; columns reach MAP only through {@link #fromSqlTypeName}. Capability registrations
+     * for filter / project operators on MAP columns are intentionally minimal: callers (e.g.
+     * PPL {@code where doc.user.name}) always wrap the MAP column in an ITEM lookup whose
+     * result type is the map's value type, so the EQUALS / sort / aggregate operators see
+     * the value-level type by the time the runtime executes them.
      */
     MAP("map");
 

sandbox/libs/analytics-framework/src/main/java/org/opensearch/analytics/spi/ScalarFunction.java

@@ -103,6 +103,15 @@ public enum ScalarFunction {
     REX_EXTRACT(Category.STRING, SqlKind.OTHER_FUNCTION),
     REX_EXTRACT_MULTI(Category.STRING, SqlKind.OTHER_FUNCTION),
     REX_OFFSET(Category.STRING, SqlKind.OTHER_FUNCTION),
+    /**
+     * PPL {@code parse <field> '<regex>'} — extracts named regex groups into a
+     * {@code MAP<VARCHAR, VARCHAR>}. Resolves by identifier-name through
+     * {@link #fromSqlFunction(SqlFunction)} ({@code SqlKind.OTHER_FUNCTION}
+     * shared with many scalar UDFs). Pairs with {@link #ITEM} downstream:
+     * {@code parse} returns the map, {@code item(map, group)} extracts each
+     * named group at the call site.
+     */
+    PARSE(Category.STRING, SqlKind.OTHER_FUNCTION),
 
     // TODO: Frontend/lang-specific functions (NUM/AUTO/MEMK/MKTIME etc.) shouldn't
     // live in the shared analytics-framework SPI — they couple the framework to PPL

sandbox/plugins/analytics-backend-datafusion/rust/Cargo.toml

@@ -69,7 +69,7 @@ serde_json = { workspace = true, features = ["preserve_order"] }
 # multi-path JSON-array output. Moving to 1.x is a follow-up once we can
 # reproduce that distinction against the new API surface.
 jsonpath-rust = "=0.7.5"
-# mvfind / rex_extract / rex_extract_multi / rex_offset UDFs — Java-Pattern-compatible
+# mvfind / parse / rex_extract / rex_extract_multi / rex_offset UDFs — Java-Pattern-compatible
 # regex matching. Same crate DataFusion's own regexp_replace / regexp_match are
 # built on, so semantics match the existing analytics-engine route's
 # `regex_match` operator (Operator::RegexMatch).

sandbox/plugins/analytics-backend-datafusion/rust/src/session_context.rs

@@ -144,10 +144,11 @@ pub async unsafe fn create_session_context(
         .build();
 
     let ctx = SessionContext::new_with_state(state);
-    // Register OpenSearch UDFs (mvappend, mvfind, mvzip, convert_tz, …) on this session
-    // so the substrait converter at execute_with_context can resolve their function names.
-    // Without this, fragment execution fails with "Unsupported function name" because
-    // df_execute_with_context reuses this handle's ctx instead of building a fresh one.
+    // Register OpenSearch UDFs (parse, item, mvappend, mvfind, mvzip, convert_tz, …)
+    // on this session so the substrait converter at execute_with_context can resolve
+    // their function names. Without this, fragment execution fails with "Unsupported
+    // function name" because df_execute_with_context reuses this handle's ctx instead
+    // of building a fresh one.
     crate::udf::register_all(&ctx);
     crate::udaf::register_all(&ctx);