pytrilogy 0.0.2.52__tar.gz → 0.0.2.54__tar.gz

{pytrilogy-0.0.2.52/pytrilogy.egg-info → pytrilogy-0.0.2.54}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: pytrilogy
-Version: 0.0.2.52
+Version: 0.0.2.54
 Summary: Declarative, typed query language that compiles to SQL.
 Home-page: 
 Author: 
@@ -34,7 +34,7 @@ Requires-Dist: snowflake-sqlalchemy; extra == "snowflake"
 
 pytrilogy is an experimental implementation of the Trilogy language, a higher-level SQL that replaces tables/joins with a lightweight semantic binding layer.
 
-Trilogy looks like SQL, but simpler. It's a modern SQL refresh targeted at SQL lovers who want reusability and simplicity with the power and iteratability of SQL. It compiles to SQL - making it easy to debug or integrate into existing workflows - and can be run against any supported SQL backend.  
+Trilogy looks like SQL, but simpler. It's a modern SQL refresh targeted at SQL lovers who want more reusability and composability without losing the expressiveness and iterative value of SQL. It compiles to SQL - making it easy to debug or integrate into existing workflows - and can be run against any supported SQL backend.  
 
 > [!TIP]
 > To get an overview of the language and run interactive examples, head to the [documentation](https://trilogydata.dev/).
@@ -193,11 +193,12 @@ address `bigquery-public-data.usa_names.usa_1910_2013`;
 executor = Dialects.BIGQUERY.default_executor(environment=environment)
 
 results = executor.execute_text(
-'''SELECT
-    name,
-    sum(yearly_name_count) -> name_count 
+'''
 WHERE
     name = 'Elvis'
+SELECT
+    name,
+    sum(yearly_name_count) -> name_count 
 ORDER BY
     name_count desc
 LIMIT 10;
@@ -270,9 +271,11 @@ Please open an issue first to discuss what you would like to change, and then cr
 ## Similar in space
 Trilogy combines two aspects; a semantic layer and a query language. Examples of both are linked below:
 
-Python "semantic layers" are tools for defining data access to a warehouse in a more abstract way.
+"semantic layers" are tools for defining an metadata layer above a SQL/warehouse base to enable higher level abstractions.
 
 - [metricsflow](https://github.com/dbt-labs/metricflow)
+- [cube](https://github.com/cube-js/cube)
+- [zillion](https://github.com/totalhack/zillion)
 
 "Better SQL" has been a popular space. We believe Trilogy takes a different approach then the following,
 but all are worth checking out. Please open PRs/comment for anything missed!
@@ -321,11 +324,14 @@ address <table>;
 Primary acces
 
 ```sql
-select
-    <concept>,
-    <concept>+1 -> <alias>
 WHERE
     <concept> = <value>
+select
+    <concept>,
+    <concept>+1 -> <alias>,
+    ...
+HAVING
+    <alias> = <value2>
 ORDER BY
     <concept> asc|desc
 ;
@@ -337,11 +343,13 @@ Reusable virtual set of rows. Useful for windows, filtering.
 
 ```sql
 with <alias> as
-select
-    <concept>,
-    <concept>+1 -> <alias>
 WHERE
     <concept> = <value>
+select
+    <concept>,
+    <concept>+1 -> <alias>,
+    ...
+
 
 select <alias>.<concept>;
 
@@ -357,6 +365,18 @@ persist <alias> as <table_name> from
 <select>;
 ```
 
+#### COPY
+
+Currently supported target types are <CSV>, though backend support may vary.
+
+```sql
+COPY INTO <TARGET_TYPE> '<target_path>' FROM SELECT
+    <concept>, ...
+ORDER BY
+    <concept>, ...
+;
+```
+
 #### SHOW
 
 Return generated SQL without executing.

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/README.md

@@ -4,7 +4,7 @@
 
 pytrilogy is an experimental implementation of the Trilogy language, a higher-level SQL that replaces tables/joins with a lightweight semantic binding layer.
 
-Trilogy looks like SQL, but simpler. It's a modern SQL refresh targeted at SQL lovers who want reusability and simplicity with the power and iteratability of SQL. It compiles to SQL - making it easy to debug or integrate into existing workflows - and can be run against any supported SQL backend.  
+Trilogy looks like SQL, but simpler. It's a modern SQL refresh targeted at SQL lovers who want more reusability and composability without losing the expressiveness and iterative value of SQL. It compiles to SQL - making it easy to debug or integrate into existing workflows - and can be run against any supported SQL backend.  
 
 > [!TIP]
 > To get an overview of the language and run interactive examples, head to the [documentation](https://trilogydata.dev/).
@@ -163,11 +163,12 @@ address `bigquery-public-data.usa_names.usa_1910_2013`;
 executor = Dialects.BIGQUERY.default_executor(environment=environment)
 
 results = executor.execute_text(
-'''SELECT
-    name,
-    sum(yearly_name_count) -> name_count 
+'''
 WHERE
     name = 'Elvis'
+SELECT
+    name,
+    sum(yearly_name_count) -> name_count 
 ORDER BY
     name_count desc
 LIMIT 10;
@@ -240,9 +241,11 @@ Please open an issue first to discuss what you would like to change, and then cr
 ## Similar in space
 Trilogy combines two aspects; a semantic layer and a query language. Examples of both are linked below:
 
-Python "semantic layers" are tools for defining data access to a warehouse in a more abstract way.
+"semantic layers" are tools for defining an metadata layer above a SQL/warehouse base to enable higher level abstractions.
 
 - [metricsflow](https://github.com/dbt-labs/metricflow)
+- [cube](https://github.com/cube-js/cube)
+- [zillion](https://github.com/totalhack/zillion)
 
 "Better SQL" has been a popular space. We believe Trilogy takes a different approach then the following,
 but all are worth checking out. Please open PRs/comment for anything missed!
@@ -291,11 +294,14 @@ address <table>;
 Primary acces
 
 ```sql
-select
-    <concept>,
-    <concept>+1 -> <alias>
 WHERE
     <concept> = <value>
+select
+    <concept>,
+    <concept>+1 -> <alias>,
+    ...
+HAVING
+    <alias> = <value2>
 ORDER BY
     <concept> asc|desc
 ;
@@ -307,11 +313,13 @@ Reusable virtual set of rows. Useful for windows, filtering.
 
 ```sql
 with <alias> as
-select
-    <concept>,
-    <concept>+1 -> <alias>
 WHERE
     <concept> = <value>
+select
+    <concept>,
+    <concept>+1 -> <alias>,
+    ...
+
 
 select <alias>.<concept>;
 
@@ -327,6 +335,18 @@ persist <alias> as <table_name> from
 <select>;
 ```
 
+#### COPY
+
+Currently supported target types are <CSV>, though backend support may vary.
+
+```sql
+COPY INTO <TARGET_TYPE> '<target_path>' FROM SELECT
+    <concept>, ...
+ORDER BY
+    <concept>, ...
+;
+```
+
 #### SHOW
 
 Return generated SQL without executing.

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54/pytrilogy.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: pytrilogy
-Version: 0.0.2.52
+Version: 0.0.2.54
 Summary: Declarative, typed query language that compiles to SQL.
 Home-page: 
 Author: 
@@ -34,7 +34,7 @@ Requires-Dist: snowflake-sqlalchemy; extra == "snowflake"
 
 pytrilogy is an experimental implementation of the Trilogy language, a higher-level SQL that replaces tables/joins with a lightweight semantic binding layer.
 
-Trilogy looks like SQL, but simpler. It's a modern SQL refresh targeted at SQL lovers who want reusability and simplicity with the power and iteratability of SQL. It compiles to SQL - making it easy to debug or integrate into existing workflows - and can be run against any supported SQL backend.  
+Trilogy looks like SQL, but simpler. It's a modern SQL refresh targeted at SQL lovers who want more reusability and composability without losing the expressiveness and iterative value of SQL. It compiles to SQL - making it easy to debug or integrate into existing workflows - and can be run against any supported SQL backend.  
 
 > [!TIP]
 > To get an overview of the language and run interactive examples, head to the [documentation](https://trilogydata.dev/).
@@ -193,11 +193,12 @@ address `bigquery-public-data.usa_names.usa_1910_2013`;
 executor = Dialects.BIGQUERY.default_executor(environment=environment)
 
 results = executor.execute_text(
-'''SELECT
-    name,
-    sum(yearly_name_count) -> name_count 
+'''
 WHERE
     name = 'Elvis'
+SELECT
+    name,
+    sum(yearly_name_count) -> name_count 
 ORDER BY
     name_count desc
 LIMIT 10;
@@ -270,9 +271,11 @@ Please open an issue first to discuss what you would like to change, and then cr
 ## Similar in space
 Trilogy combines two aspects; a semantic layer and a query language. Examples of both are linked below:
 
-Python "semantic layers" are tools for defining data access to a warehouse in a more abstract way.
+"semantic layers" are tools for defining an metadata layer above a SQL/warehouse base to enable higher level abstractions.
 
 - [metricsflow](https://github.com/dbt-labs/metricflow)
+- [cube](https://github.com/cube-js/cube)
+- [zillion](https://github.com/totalhack/zillion)
 
 "Better SQL" has been a popular space. We believe Trilogy takes a different approach then the following,
 but all are worth checking out. Please open PRs/comment for anything missed!
@@ -321,11 +324,14 @@ address <table>;
 Primary acces
 
 ```sql
-select
-    <concept>,
-    <concept>+1 -> <alias>
 WHERE
     <concept> = <value>
+select
+    <concept>,
+    <concept>+1 -> <alias>,
+    ...
+HAVING
+    <alias> = <value2>
 ORDER BY
     <concept> asc|desc
 ;
@@ -337,11 +343,13 @@ Reusable virtual set of rows. Useful for windows, filtering.
 
 ```sql
 with <alias> as
-select
-    <concept>,
-    <concept>+1 -> <alias>
 WHERE
     <concept> = <value>
+select
+    <concept>,
+    <concept>+1 -> <alias>,
+    ...
+
 
 select <alias>.<concept>;
 
@@ -357,6 +365,18 @@ persist <alias> as <table_name> from
 <select>;
 ```
 
+#### COPY
+
+Currently supported target types are <CSV>, though backend support may vary.
+
+```sql
+COPY INTO <TARGET_TYPE> '<target_path>' FROM SELECT
+    <concept>, ...
+ORDER BY
+    <concept>, ...
+;
+```
+
 #### SHOW
 
 Return generated SQL without executing.

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/tests/test_parsing.py

@@ -182,7 +182,7 @@ select
     ]:
         assert f"local.{name}" in env.concepts
         assert env.concepts[name].purpose == Purpose.PROPERTY
-        assert env.concepts[name].keys == (env.concepts["id"],)
+        assert env.concepts[name].keys == {env.concepts["id"].address}
 
 
 def test_purpose_and_derivation():
@@ -200,10 +200,10 @@ select
 
     for name in ["join_id"]:
         assert env.concepts[name].purpose == Purpose.PROPERTY
-        assert env.concepts[name].keys == (
-            env.concepts["id"],
-            env.concepts["other_id"],
-        )
+        assert env.concepts[name].keys == {
+            env.concepts["id"].address,
+            env.concepts["other_id"].address,
+        }
 
 
 def test_output_purpose():

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/tests/test_partial_handling.py

@@ -35,7 +35,9 @@ def setup_titanic(env: Environment):
         namespace=namespace,
         datatype=DataType.INTEGER,
         purpose=Purpose.PROPERTY,
-        keys=(id,),
+        keys={
+            id.address,
+        },
     )
 
     name = Concept(
@@ -43,7 +45,9 @@ def setup_titanic(env: Environment):
         namespace=namespace,
         datatype=DataType.STRING,
         purpose=Purpose.PROPERTY,
-        keys=(id,),
+        keys={
+            id.address,
+        },
     )
 
     pclass = Concept(
@@ -51,21 +55,27 @@ def setup_titanic(env: Environment):
         namespace=namespace,
         purpose=Purpose.PROPERTY,
         datatype=DataType.INTEGER,
-        keys=(id,),
+        keys={
+            id.address,
+        },
     )
     survived = Concept(
         name="survived",
         namespace=namespace,
         purpose=Purpose.PROPERTY,
         datatype=DataType.INTEGER,
-        keys=(id,),
+        keys={
+            id.address,
+        },
     )
     fare = Concept(
         name="fare",
         namespace=namespace,
         purpose=Purpose.PROPERTY,
         datatype=DataType.FLOAT,
-        keys=(id,),
+        keys={
+            id.address,
+        },
     )
     for x in [id, age, survived, name, pclass, fare]:
         env.add_concept(x)

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/trilogy/__init__.py

@@ -4,6 +4,6 @@ from trilogy.dialect.enums import Dialects
 from trilogy.executor import Executor
 from trilogy.parser import parse
 
-__version__ = "0.0.2.52"
+__version__ = "0.0.2.54"
 
 __all__ = ["parse", "Executor", "Dialects", "Environment", "CONFIG"]

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/trilogy/core/environment_helpers.py

@@ -50,7 +50,9 @@ def generate_date_concepts(concept: Concept, environment: Environment):
             lineage=const_function,
             grain=const_function.output_grain,
             namespace=namespace,
-            keys=(concept,),
+            keys=set(
+                concept.address,
+            ),
             metadata=Metadata(
                 description=f"Auto-derived. Integer format. The {ftype.value} derived from {concept.name}, {base_description}",
                 line_number=base_line_number,
@@ -99,7 +101,9 @@ def generate_datetime_concepts(concept: Concept, environment: Environment):
             lineage=const_function,
             grain=const_function.output_grain,
             namespace=namespace,
-            keys=(concept,),
+            keys=set(
+                concept.address,
+            ),
             metadata=Metadata(
                 description=f"Auto-derived. Integer format. The {ftype.value} derived from {concept.name}, {base_description}",
                 line_number=base_line_number,
@@ -139,7 +143,9 @@ def generate_key_concepts(concept: Concept, environment: Environment):
             lineage=const_function,
             grain=const_function.output_grain,
             namespace=namespace,
-            keys=(concept,),
+            keys={
+                concept.address,
+            },
             metadata=Metadata(
                 description=f"Auto-derived. Integer format. The {ftype.value} derived from {concept.name}, {base_description}",
                 line_number=base_line_number,

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/trilogy/core/models.py

@@ -427,12 +427,15 @@ class Concept(Mergeable, Namespaced, SelectContext, BaseModel):
         ]
     ] = None
     namespace: Optional[str] = Field(default=DEFAULT_NAMESPACE, validate_default=True)
-    keys: Optional[Tuple["Concept", ...]] = None
+    keys: Optional[set[str]] = None
     grain: "Grain" = Field(default=None, validate_default=True)  # type: ignore
     modifiers: List[Modifier] = Field(default_factory=list)  # type: ignore
     pseudonyms: set[str] = Field(default_factory=set)
     _address_cache: str | None = None
 
+    def __init__(self, **data):
+        super().__init__(**data)
+
     def duplicate(self) -> Concept:
         return self.model_copy(deep=True)
 
@@ -480,7 +483,7 @@ class Concept(Mergeable, Namespaced, SelectContext, BaseModel):
             grain=self.grain.with_merge(source, target, modifiers),
             namespace=self.namespace,
             keys=(
-                tuple(x.with_merge(source, target, modifiers) for x in self.keys)
+                set(x if x != source.address else target.address for x in self.keys)
                 if self.keys
                 else None
             ),
@@ -488,17 +491,6 @@ class Concept(Mergeable, Namespaced, SelectContext, BaseModel):
             pseudonyms=self.pseudonyms,
         )
 
-    @field_validator("keys", mode="before")
-    @classmethod
-    def keys_validator(cls, v, info: ValidationInfo):
-        if v is None:
-            return v
-        if not isinstance(v, (list, tuple)):
-            raise ValueError(f"Keys must be a list or tuple, got {type(v)}")
-        if isinstance(v, list):
-            return tuple(v)
-        return v
-
     @field_validator("namespace", mode="plain")
     @classmethod
     def namespace_validation(cls, v):
@@ -610,7 +602,7 @@ class Concept(Mergeable, Namespaced, SelectContext, BaseModel):
                 else namespace
             ),
             keys=(
-                tuple([x.with_namespace(namespace) for x in self.keys])
+                set([address_with_namespace(x, namespace) for x in self.keys])
                 if self.keys
                 else None
             ),
@@ -627,25 +619,17 @@ class Concept(Mergeable, Namespaced, SelectContext, BaseModel):
                 local_concepts=local_concepts, grain=grain, environment=environment
             )
         final_grain = self.grain or grain
-        keys = (
-            tuple(
-                [
-                    x.with_select_context(local_concepts, grain, environment)
-                    for x in self.keys
-                ]
-            )
-            if self.keys
-            else None
-        )
+        keys = self.keys if self.keys else None
         if self.is_aggregate and isinstance(new_lineage, Function):
             grain_components = [environment.concepts[c] for c in grain.components]
             new_lineage = AggregateWrapper(function=new_lineage, by=grain_components)
             final_grain = grain
-            keys = tuple(grain_components)
+            keys = set(grain.components)
         elif (
             self.is_aggregate and not keys and isinstance(new_lineage, AggregateWrapper)
         ):
-            keys = tuple(new_lineage.by)
+            keys = set([x.address for x in new_lineage.by])
+
         return self.__class__(
             name=self.name,
             datatype=self.datatype,
@@ -687,13 +671,10 @@ class Concept(Mergeable, Namespaced, SelectContext, BaseModel):
             if self.lineage:
                 for item in self.lineage.arguments:
                     if isinstance(item, Concept):
-                        if item.keys and not all(c in components for c in item.keys):
-                            components += item.sources
-                        else:
-                            components += item.sources
+                        components += [x.address for x in item.sources]
             # TODO: set synonyms
             grain = Grain(
-                components=set([x.address for x in components]),
+                components=set([x for x in components]),
             )  # synonym_set=generate_concept_synonyms(components))
         elif self.purpose == Purpose.METRIC:
             grain = Grain()
@@ -1058,13 +1039,19 @@ class EnvironmentConceptDict(dict):
             if DEFAULT_NAMESPACE + "." + key in self:
                 return self.__getitem__(DEFAULT_NAMESPACE + "." + key, line_no)
             if not self.fail_on_missing:
+                if "." in key:
+                    ns, rest = key.rsplit(".", 1)
+                else:
+                    ns = DEFAULT_NAMESPACE
+                    rest = key
                 if key in self.undefined:
                     return self.undefined[key]
                 undefined = UndefinedConcept(
-                    name=key,
+                    name=rest,
                     line_no=line_no,
                     datatype=DataType.UNKNOWN,
                     purpose=Purpose.UNKNOWN,
+                    namespace=ns,
                 )
                 self.undefined[key] = undefined
                 return undefined
@@ -1365,18 +1352,6 @@ class Function(Mergeable, Namespaced, SelectContext, BaseModel):
             base_grain += input.grain
         return base_grain
 
-    @property
-    def output_keys(self) -> list[Concept]:
-        # aggregates have an abstract grain
-        components = []
-        # scalars have implicit grain of all arguments
-        for input in self.concept_arguments:
-            if input.purpose == Purpose.KEY:
-                components.append(input)
-            elif input.keys:
-                components += input.keys
-        return list(set(components))
-
 
 class ConceptTransform(Namespaced, BaseModel):
     function: Function | FilterItem | WindowItem | AggregateWrapper
@@ -3203,7 +3178,9 @@ class UndefinedConcept(Concept, Mergeable, Namespaced):
             rval = local_concepts[self.address]
             rval = rval.with_select_context(local_concepts, grain, environment)
             return rval
-        environment.concepts.raise_undefined(self.address, line_no=self.line_no)
+        if environment.concepts.fail_on_missing:
+            environment.concepts.raise_undefined(self.address, line_no=self.line_no)
+        return self
 
 
 class EnvironmentDatasourceDict(dict):
@@ -3336,6 +3313,8 @@ class Environment(BaseModel):
 
     @classmethod
     def from_file(cls, path: str | Path) -> "Environment":
+        if isinstance(path, str):
+            path = Path(path)
         with open(path, "r") as f:
             read = f.read()
         return Environment(working_path=Path(path).parent).parse(read)[0]
@@ -3645,11 +3624,6 @@ class Environment(BaseModel):
                     self.merge_concept(new_concept, current_concept, [])
                 else:
                     self.add_concept(current_concept, meta=meta, _ignore_cache=True)
-
-            # else:
-            #     self.add_concept(
-            #         current_concept, meta=meta, _ignore_cache=True
-            #     )
         if not _ignore_cache:
             self.gen_concept_list_caches()
         return datasource
@@ -4521,13 +4495,11 @@ class RowsetDerivationStatement(HasUUID, Namespaced, BaseModel):
         # remap everything to the properties of the rowset
         for x in output:
             if x.keys:
-                if all([k.address in orig for k in x.keys]):
-                    x.keys = tuple(
-                        [orig[k.address] if k.address in orig else k for k in x.keys]
-                    )
+                if all([k in orig for k in x.keys]):
+                    x.keys = set([orig[k].address if k in orig else k for k in x.keys])
                 else:
                     # TODO: fix this up
-                    x.keys = tuple()
+                    x.keys = set()
         for x in output:
             if all([c in orig for c in x.grain.components]):
                 x.grain = Grain(

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/trilogy/core/processing/node_generators/common.py

@@ -26,6 +26,7 @@ def resolve_function_parent_concepts(
     if not isinstance(concept.lineage, (Function, AggregateWrapper)):
         raise ValueError(f"Concept {concept} lineage is not function or aggregate")
     if concept.derivation == PurposeLineage.AGGREGATE:
+        base: list[Concept] = []
         if not concept.grain.abstract:
             base = concept.lineage.concept_arguments + [
                 environment.concepts[c] for c in concept.grain.components
@@ -41,7 +42,7 @@ def resolve_function_parent_concepts(
             extra_property_grain = concept.lineage.concept_arguments
         for x in extra_property_grain:
             if isinstance(x, Concept) and x.purpose == Purpose.PROPERTY and x.keys:
-                base += x.keys
+                base += [environment.concepts[c] for c in x.keys]
         return unique(base, "address")
     # TODO: handle basic lineage chains?
     return unique(concept.lineage.concept_arguments, "address")
@@ -81,7 +82,7 @@ def resolve_filter_parent_concepts(
         and direct_parent.purpose == Purpose.PROPERTY
         and direct_parent.keys
     ):
-        base_rows += direct_parent.keys
+        base_rows += [environment.concepts[c] for c in direct_parent.keys]
     if concept.lineage.where.existence_arguments:
         return (
             concept.lineage.content,
@@ -106,7 +107,7 @@ def gen_property_enrichment_node(
     for x in extra_properties:
         if not x.keys:
             raise SyntaxError(f"Property {x.address} missing keys in lookup")
-        keys = "-".join([y.address for y in x.keys])
+        keys = "-".join([y for y in x.keys])
         required_keys[keys].add(x.address)
     final_nodes = []
     for _k, vs in required_keys.items():

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/trilogy/core/processing/node_generators/filter_node.py

@@ -138,7 +138,7 @@ def gen_filter_node(
         )
         parent.grain = Grain.from_concepts(
             (
-                list(immediate_parent.keys)
+                [environment.concepts[k] for k in immediate_parent.keys]
                 if immediate_parent.keys
                 else [immediate_parent]
             )
@@ -146,7 +146,8 @@ def gen_filter_node(
                 x
                 for x in local_optional
                 if x.address in [y.address for y in parent.output_concepts]
-            ]
+            ],
+            environment=environment,
         )
         parent.rebuild_cache()
         filter_node = parent

{pytrilogy-0.0.2.52 → pytrilogy-0.0.2.54}/trilogy/core/processing/node_generators/rowset_node.py

@@ -97,7 +97,6 @@ def gen_rowset_node(
     possible_joins = concept_to_relevant_joins(
         [x for x in node.output_concepts if x.derivation != PurposeLineage.ROWSET]
     )
-    logger.info({x.address: x.keys for x in possible_joins})
     if not possible_joins:
         logger.info(
             f"{padding(depth)}{LOGGER_PREFIX} no possible joins for rowset node to get {[x.address for x in local_optional]}; have {[x.address for x in node.output_concepts]}"