tinybird 0.0.1.dev178__tar.gz → 0.0.1.dev180__tar.gz

{tinybird-0.0.1.dev178 → tinybird-0.0.1.dev180}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: tinybird
-Version: 0.0.1.dev178
+Version: 0.0.1.dev180
 Summary: Tinybird Command Line Tool
 Home-page: https://www.tinybird.co/docs/forward/commands
 Author: Tinybird

{tinybird-0.0.1.dev178 → tinybird-0.0.1.dev180}/tinybird/sql_toolset.py

@@ -1,5 +1,6 @@
 import copy
 import logging
+import re
 from collections import defaultdict
 from datetime import datetime
 from functools import lru_cache
@@ -42,10 +43,88 @@ class InvalidResource(ValueError):
         self.table = table
 
 
+class UnoptimizedJoinException(Exception):
+    def __init__(self, sql: str):
+        self.sql = sql
+        self.msg = f"Materialized node SQL contains a join that is not optimized: {sql}"
+        self.documentation = (
+            "/docs/work-with-data/optimization/opt201-fix-mistakes#understanding-the-materialized-join-issue"
+        )
+        super().__init__(self.msg)
+
+
+ChQueryTable = Tuple[Optional[str], Optional[str], Optional[str]]
+
+
+def get_left_table(sql: str, default_database: Optional[str] = None) -> ChQueryTable:
+    if default_database is None:
+        left_table = chquery.get_left_table(sql)
+    else:
+        left_table = chquery.get_left_table(sql, default_database=default_database)
+    return left_table
+
+
 def format_sql(sql: str) -> str:
     return chquery.format(sql)
 
 
+def explain_plan(sql: str) -> str:
+    return chquery.explain_ast(sql)
+
+
+def has_join(sql: str) -> bool:
+    return any(line.rstrip().startswith("TableJoin") for line in explain_plan(sql).split())
+
+
+def has_unoptimized_join(sql: str, left_table: Optional[Union[Tuple[str, str], Tuple[str, str, str]]] = None) -> None:
+    """
+    Check if a SQL query contains an unoptimized join.
+    A join is considered optimized if the right table is filtered by the left table's data.
+
+    Args:
+        sql: The SQL query to check
+        left_table: Optional tuple of (database, table) for the left table
+
+    Raises:
+        UnoptimizedJoin: If an unoptimized join is found
+    """
+    # TODO: We should check that we are filtering the right table by the left table's data
+    # TODO: We should check if using EXPLAIN AST is better than using regex
+
+    number_of_joins = sum(1 for line in explain_plan(sql).split() if line.rstrip().startswith("TableJoin"))
+    if number_of_joins == 0:
+        return
+
+    if not left_table:
+        left_table = chquery.get_left_table(sql)
+        if not left_table:
+            return
+
+    # Find all JOIN clauses with subqueries
+    # This pattern matches anything between JOIN and ON/USING
+    join_pattern = r"(?:LEFT\s+|RIGHT\s+|INNER\s+|FULL\s+OUTER\s+)?JOIN\s*\((.*?)\)\s+(?:AS\s+\w+)?\s*(?:ON|USING)"
+
+    # Find all joins with subqueries
+    join_matches = list(re.finditer(join_pattern, sql, re.IGNORECASE | re.DOTALL))
+
+    if number_of_joins != len(join_matches):
+        logging.debug(f"number_of_joins: {number_of_joins}, join_matches: {join_matches}")
+        raise UnoptimizedJoinException(sql)
+
+    # If no joins with subqueries found, probably is an unoptimized join
+    if not join_matches:
+        raise UnoptimizedJoinException(sql)
+
+    # Check if the left table is referenced in the subquery
+    left_table_ref = f"{left_table[0]}.{left_table[1]}"
+
+    for match in join_matches:
+        subquery = match.group(1)  # Get the captured subquery
+        logging.debug(f"subquery: {subquery} left_table_ref: {left_table_ref}")
+        if left_table_ref not in subquery:
+            raise UnoptimizedJoinException(sql)
+
+
 def format_where_for_mutation_command(where_clause: str) -> str:
     """
     >>> format_where_for_mutation_command("numnights = 99")

{tinybird-0.0.1.dev178 → tinybird-0.0.1.dev180}/tinybird/tb/__cli__.py

@@ -4,5 +4,5 @@ __description__ = 'Tinybird Command Line Tool'
 __url__ = 'https://www.tinybird.co/docs/forward/commands'
 __author__ = 'Tinybird'
 __author_email__ = 'support@tinybird.co'
-__version__ = '0.0.1.dev178'
-__revision__ = 'c2632b6'
+__version__ = '0.0.1.dev180'
+__revision__ = '4ea1a7e'

{tinybird-0.0.1.dev178 → tinybird-0.0.1.dev180}/tinybird/tb/modules/build.py

@@ -89,7 +89,6 @@ def build_project(project: Project, tb_client: TinyB, silent: bool = False) -> O
         files = [
             ("context://", ("cli-version", "1.0.0", "text/plain")),
         ]
-        fds = []
         project_path = project.path
         project_files = project.get_project_files()
 
@@ -98,10 +97,11 @@ def build_project(project: Project, tb_client: TinyB, silent: bool = False) -> O
 
         for file_path in project_files:
             relative_path = str(Path(file_path).relative_to(project_path))
-            fd = open(file_path, "rb")
-            fds.append(fd)
-            content_type = DATAFILE_TYPE_TO_CONTENT_TYPE.get(Path(file_path).suffix, "application/unknown")
-            files.append((MULTIPART_BOUNDARY_DATA_PROJECT, (relative_path, fd.read().decode("utf-8"), content_type)))
+            with open(file_path, "rb") as fd:
+                content_type = DATAFILE_TYPE_TO_CONTENT_TYPE.get(Path(file_path).suffix, "application/unknown")
+                files.append(
+                    (MULTIPART_BOUNDARY_DATA_PROJECT, (relative_path, fd.read().decode("utf-8"), content_type))
+                )
         HEADERS = {"Authorization": f"Bearer {TINYBIRD_API_KEY}"}
 
         r = requests.post(TINYBIRD_API_URL, files=files, headers=HEADERS)
@@ -195,9 +195,6 @@ def build_project(project: Project, tb_client: TinyB, silent: bool = False) -> O
             error = f"Unknown build result. Error: {result.get('error')}"
     except Exception as e:
         error = str(e)
-    finally:
-        for fd in fds:
-            fd.close()
 
     if error:
         raise click.ClickException(error)

{tinybird-0.0.1.dev178 → tinybird-0.0.1.dev180}/tinybird/tb/modules/deployment.py

@@ -524,13 +524,11 @@ def create_deployment(
     files = [
         ("context://", ("cli-version", "1.0.0", "text/plain")),
     ]
-    fds = []
     for file_path in project.get_project_files():
         relative_path = str(Path(file_path).relative_to(project.path))
-        fd = open(file_path, "rb")
-        fds.append(fd)
-        content_type = DATAFILE_TYPE_TO_CONTENT_TYPE.get(Path(file_path).suffix, "application/unknown")
-        files.append((MULTIPART_BOUNDARY_DATA_PROJECT, (relative_path, fd.read().decode("utf-8"), content_type)))
+        with open(file_path, "rb") as fd:
+            content_type = DATAFILE_TYPE_TO_CONTENT_TYPE.get(Path(file_path).suffix, "application/unknown")
+            files.append((MULTIPART_BOUNDARY_DATA_PROJECT, (relative_path, fd.read().decode("utf-8"), content_type)))
 
     deployment = None
     try:
@@ -605,9 +603,6 @@ def create_deployment(
             click.echo(FeedbackManager.error(message=f"Unknown deployment result {status}"))
     except Exception as e:
         click.echo(FeedbackManager.error_exception(error=e))
-    finally:
-        for fd in fds:
-            fd.close()
 
         if not deployment and not check:
             sys_exit("deployment_error", "Deployment failed")

{tinybird-0.0.1.dev178 → tinybird-0.0.1.dev180}/tinybird.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: tinybird
-Version: 0.0.1.dev178
+Version: 0.0.1.dev180
 Summary: Tinybird Command Line Tool
 Home-page: https://www.tinybird.co/docs/forward/commands
 Author: Tinybird