check index schema for matching support

agourlay · agourlay · commit f5b33608179a · 2025-04-30T10:20:41.000+02:00
diff --git a/lib/collection/src/collection/payload_index_schema.rs b/lib/collection/src/collection/payload_index_schema.rs
@@ -97,8 +97,8 @@ impl Collection {
         Ok(result)
     }
 
-    pub fn payload_key_is_indexed(&self, key: &JsonPath) -> bool {
-        self.payload_index_schema.read().schema.contains_key(key)
+    pub fn payload_key_index_schema(&self, key: &JsonPath) -> Option<PayloadFieldSchema> {
+        self.payload_index_schema.read().schema.get(key).cloned()
     }
 
     /// Returns an arbitrary payload key along with acceptable
diff --git a/lib/collection/src/operations/verification/mod.rs b/lib/collection/src/operations/verification/mod.rs
@@ -11,6 +11,7 @@ mod update;
 
 use std::fmt::Display;
 
+use segment::json_path::JsonPath;
 use segment::types::{Filter, SearchParams, StrictModeConfig};
 
 use super::types::{CollectionError, CollectionResult};
@@ -295,6 +296,32 @@ impl StrictModeVerification for SearchParams {
     }
 }
 
+pub fn check_grouping_field(
+    group_by: &JsonPath,
+    collection: &Collection,
+    strict_mode_config: &StrictModeConfig,
+) -> CollectionResult<()> {
+    // check for unindexed fields targeted by group_by
+    if strict_mode_config.unindexed_filtering_retrieve == Some(false) {
+        // check the group_by field is indexed and support `match` statement
+        if let Some(schema) = collection.payload_key_index_schema(group_by) {
+            let schema_kind = schema.kind();
+            if !schema_kind.support_match() {
+                return Err(CollectionError::strict_mode(
+                    format!("Index of type \"{schema_kind:?}\" found for \"{group_by}\""),
+                    "Create an index supporting `match` for this key.",
+                ));
+            }
+        } else {
+            return Err(CollectionError::strict_mode(
+                format!("Index required but not found for \"{group_by}\""),
+                "Create an index supporting `match` for this key.",
+            ));
+        }
+    }
+    Ok(())
+}
+
 #[cfg(test)]
 mod test {
     use std::sync::Arc;
diff --git a/lib/collection/src/operations/verification/query.rs b/lib/collection/src/operations/verification/query.rs
@@ -1,6 +1,6 @@
 use segment::types::StrictModeConfig;
 
-use super::StrictModeVerification;
+use super::{StrictModeVerification, check_grouping_field};
 use crate::collection::Collection;
 use crate::operations::types::{CollectionError, CollectionResult};
 use crate::operations::universal_query::collection_query::{
@@ -128,14 +128,7 @@ impl StrictModeVerification for CollectionQueryGroupsRequest {
             query.check_strict_mode(collection, strict_mode_config)?
         }
         // check for unindexed fields targeted by group_by
-        if strict_mode_config.unindexed_filtering_retrieve == Some(false)
-            && !collection.payload_key_is_indexed(&self.group_by)
-        {
-            return Err(CollectionError::strict_mode(
-                format!("Index required but not found for \"{}\"", self.group_by,),
-                "Create an index for this key.",
-            ));
-        }
+        check_grouping_field(&self.group_by, collection, strict_mode_config)?;
         Ok(())
     }
 
diff --git a/lib/collection/src/operations/verification/search.rs b/lib/collection/src/operations/verification/search.rs
@@ -1,9 +1,11 @@
 use api::rest::{SearchGroupsRequestInternal, SearchRequestInternal};
 use segment::types::{Filter, SearchParams, StrictModeConfig};
 
-use super::StrictModeVerification;
+use super::{StrictModeVerification, check_grouping_field};
 use crate::collection::Collection;
-use crate::operations::types::{CollectionError, CollectionResult, CoreSearchRequest, SearchRequestBatch};
+use crate::operations::types::{
+    CollectionError, CollectionResult, CoreSearchRequest, SearchRequestBatch,
+};
 
 impl StrictModeVerification for SearchRequestInternal {
     fn indexed_filter_read(&self) -> Option<&Filter> {
@@ -92,17 +94,7 @@ impl StrictModeVerification for SearchGroupsRequestInternal {
         strict_mode_config: &StrictModeConfig,
     ) -> CollectionResult<()> {
         // check for unindexed fields targeted by group_by
-        if strict_mode_config.unindexed_filtering_retrieve == Some(false)
-            && !collection.payload_key_is_indexed(&self.group_request.group_by)
-        {
-            return Err(CollectionError::strict_mode(
-                format!(
-                    "Index required but not found for \"{}\"",
-                    self.group_request.group_by
-                ),
-                "Create an index for this key.",
-            ));
-        }
+        check_grouping_field(&self.group_request.group_by, collection, strict_mode_config)?;
         Ok(())
     }
 
diff --git a/lib/segment/src/types.rs b/lib/segment/src/types.rs
@@ -1711,6 +1711,20 @@ impl PayloadSchemaType {
             Self::Uuid => PayloadSchemaParams::Uuid(UuidIndexParams::default()),
         }
     }
+
+    /// Check if this type supports a `match` condition
+    pub fn support_match(&self) -> bool {
+        match self {
+            Self::Keyword => true,
+            Self::Integer => true,
+            Self::Bool => true,
+            Self::Uuid => true,
+            Self::Float => false,
+            Self::Geo => false,
+            Self::Text => false,
+            Self::Datetime => false,
+        }
+    }
 }
 
 /// Payload type with parameters
diff --git a/tests/openapi/test_strictmode.py b/tests/openapi/test_strictmode.py
@@ -1860,7 +1860,7 @@ def test_strict_mode_group_by_unindexed(collection_name):
     )
 
     assert not response.ok
-    assert "Forbidden: Index required but not found for \"docId\". Help: Create an index for this key." in response.json()['status']['error']
+    assert "Forbidden: Index required but not found for \"docId\". Help: Create an index supporting `match` for this key." in response.json()['status']['error']
 
     response = request_with_validation(
         api="/collections/{collection_name}/points/query/groups",
@@ -1875,9 +1875,53 @@ def test_strict_mode_group_by_unindexed(collection_name):
         },
     )
     assert not response.ok
-    assert "Forbidden: Index required but not found for \"docId\". Help: Create an index for this key." in response.json()['status']['error']
+    assert "Forbidden: Index required but not found for \"docId\". Help: Create an index supporting `match` for this key." in response.json()['status']['error']
 
-    # create index
+    # create geo index
+    request_with_validation(
+        api='/collections/{collection_name}/index',
+        method="PUT",
+        path_params={'collection_name': collection_name},
+        query_params={'wait': 'true'},
+        body={
+            "field_name": "docId",
+            "field_schema": "geo"
+        }
+    ).raise_for_status()
+
+    # try again with geo index
+    response = request_with_validation(
+        api="/collections/{collection_name}/points/search/groups",
+        method="POST",
+        path_params={"collection_name": collection_name},
+        body={
+            "vector": [1.0, 0.0, 0.0, 0.0],
+            "limit": 10,
+            "with_payload": True,
+            "group_by": "docId",
+            "group_size": 3,
+        },
+    )
+
+    assert not response.ok
+    assert "Forbidden: Index of type \"Geo\" found for \"docId\". Help: Create an index supporting `match` for this key." in response.json()['status']['error']
+
+    response = request_with_validation(
+        api="/collections/{collection_name}/points/query/groups",
+        method="POST",
+        path_params={"collection_name": collection_name},
+        body={
+            "query": [1.0, 0.0, 0.0, 0.0],
+            "limit": 10,
+            "with_payload": True,
+            "group_by": "docId",
+            "group_size": 3,
+        },
+    )
+    assert not response.ok
+    assert "Forbidden: Index of type \"Geo\" found for \"docId\". Help: Create an index supporting `match` for this key." in response.json()['status']['error']
+
+    # create keyword index (supporting match)
     request_with_validation(
         api='/collections/{collection_name}/index',
         method="PUT",

Original file line number	Diff line number	Diff line change
`@@ -97,8 +97,8 @@ impl Collection {`
`97`	`97`	`Ok(result)`
`98`	`98`	`}`
`99`	`99`
`100`		`- pub fn payload_key_is_indexed(&self, key: &JsonPath) -> bool {`
`101`		`- self.payload_index_schema.read().schema.contains_key(key)`
	`100`	`+ pub fn payload_key_index_schema(&self, key: &JsonPath) -> Option<PayloadFieldSchema> {`
	`101`	`+ self.payload_index_schema.read().schema.get(key).cloned()`
`102`	`102`	`}`
`103`	`103`
`104`	`104`	`/// Returns an arbitrary payload key along with acceptable`