sqlh 0.3.1__tar.gz → 0.3.2__tar.gz

{sqlh-0.3.1 → sqlh-0.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: sqlh
-Version: 0.3.1
+Version: 0.3.2
 Summary: A lightweight SQL lineage analysis library for tracking table dependencies in data pipelines
 Keywords: sql,lineage,data-pipeline,dag,dependency,database,etl,data-engineering
 Requires-Python: >=3.10

{sqlh-0.3.1 → sqlh-0.3.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "sqlh"
-version = "0.3.1"
+version = "0.3.2"
 
 description = "A lightweight SQL lineage analysis library for tracking table dependencies in data pipelines"
 readme = "README.md"

{sqlh-0.3.1 → sqlh-0.3.2}/sqlh/__init__.py

@@ -14,7 +14,7 @@ from .utils import (
     table_count
 )
 
-__version__ = "0.3.1"
+__version__ = "0.3.2"
 
 __all__ = [
     "split_sql",

{sqlh-0.3.1 → sqlh-0.3.2}/sqlh/core/helper.py

@@ -444,7 +444,7 @@ def get_source_target_tables(sql: str) -> dict[str, list[str]] | None:
         ParseException: If SQL contains multiple statements
 
     Note:
-        TODO: 不能识别join后面的 [hint] table_name（已在 get_source_target_tables_v2 实现）
+        TODO: 不能识别join后面的 [hint] table_name
         {
             "source_tables": [(t1, 1), (t2, 2), (t3, 3)],
             "target_tables": [(t4, 1)]
@@ -563,158 +563,6 @@ def get_source_target_tables(sql: str) -> dict[str, list[str]] | None:
         return
 
 
-def get_source_target_tables_v2(sql: str) -> dict[str, list[str]] | None:
-    """
-    Extract source and target tables from a single SQL statement using a token-based approach.
-    changelog: 
-        -  支持识别join后面的 [hint] table_name
-    """
-    # 1. 预处理
-    clean_sql = trim_comment(sql).strip()
-    if not clean_sql:
-        return None
-
-    # 去除末尾分号
-    if clean_sql.endswith(";"):
-        clean_sql = clean_sql[:-1]
-
-    # 校验多语句
-    # 注意：这里直接传原始 sql 给 split_sql 可能更稳妥，或者传 clean_sql 取决于 split_sql 的实现
-    if len(split_sql(clean_sql)) > 1:
-        raise ParseException("sql脚本为多条SQL语句,需传入单条SQL语句.")
-
-    # 2. 分词 (使用正则优化性能)
-    # 匹配单词、括号、逗号。忽略空白字符。
-    # tokens = re.findall(r"[A-Za-z_][A-Za-z0-9_]*|[()]|,", clean_sql)
-    tokens = re.findall(r"\[[^\]]+\]|[A-Za-z_][A-Za-z0-9_]*|[()]|,", clean_sql)
-
-    if not tokens:
-        return None
-
-    # 3. 状态机变量
-    source_tables = set()
-    target_tables = set()
-
-    # 状态标志
-    # 状态标志
-    is_insert_context = False
-    is_from_context = False
-    is_join_context = False  # 新增：专门标记 JOIN 后的状态
-    is_using_context = False
-    is_merge_context = False
-
-    # 辅助变量
-
-    for token in tokens:
-        token_upper = token.upper()
-
-        # --- 特殊处理：跳过 Hint 内容 ---
-        # 如果 token 是 [...] 格式，直接跳过，不要把它当成括号或关键字
-        if token.startswith("[") and token.endswith("]"):
-            # 如果是在 JOIN 后面遇到的 Hint，保持 join_context 为 True
-            # 如果是在 FROM 后面遇到的 Hint，保持 from_context 为 True
-            continue
-
-        # --- 上下文重置 ---
-        if token == "(":
-            # 左括号通常意味着子查询，重置 FROM/JOIN 上下文
-            # 但我们不能重置 INSERT 上下文
-            is_from_context = False
-            is_join_context = False
-            is_using_context = False
-            continue
-
-        # --- 关键字状态流转 ---
-
-        if token_upper in KeyWords.insert_keywords:
-            is_insert_context = True
-            is_from_context = False
-            is_join_context = False
-            continue
-
-        if token_upper == "FROM":
-            is_from_context = True
-            is_join_context = False
-            is_insert_context = False
-            continue
-
-        # 修复核心：处理 JOIN
-        if token_upper in ["JOIN", "INNER", "LEFT", "RIGHT", "OUTER", "CROSS"]:
-            is_from_context = True  # JOIN 也是一种来源
-            is_join_context = True  # 标记刚刚遇到了 JOIN，下一个非关键字即为表名
-            continue
-
-        # 【关键修复】：遇到 ON，说明表名部分结束了，后面是关联条件
-        if token_upper == "ON":
-            is_from_context = False
-            is_join_context = False
-            continue
-
-        if token_upper == "MERGE":
-            is_merge_context = True
-            continue
-
-        if token_upper == "USING":
-            is_from_context = False
-            is_join_context = False
-            if is_merge_context:
-                is_using_context = True
-            continue
-
-        # --- 表名捕获逻辑 ---
-
-        # 1. INSERT 目标表
-        if is_insert_context and token_upper not in KeyWords.keywords and token not in ("(", ")"):
-            target_tables.add(token)
-            is_insert_context = False
-            continue
-
-        # 2. MERGE 逻辑 (略，保持原有逻辑) ...
-        if is_merge_context and not is_using_context and token_upper not in KeyWords.keywords:
-            target_tables.add(token)
-            is_merge_context = False
-            continue
-
-        if is_using_context and token_upper not in KeyWords.keywords:
-            if token != "(":
-                source_tables.add(token)
-            is_using_context = False
-            is_merge_context = False
-            continue
-
-        # 3. FROM / JOIN 源表 (修复核心)
-        # 条件：处于 FROM 或 JOIN 状态，且不是关键字，且不是括号/逗号
-        if (is_from_context or is_join_context) and token_upper not in KeyWords.keywords and token not in ("(", ")", ","):
-            source_tables.add(token)
-
-            # 捕获到表名后，重置状态，等待下一个 JOIN 或逗号
-            is_join_context = False
-            # 注意：is_from_context 保持 True，以便处理 "FROM t1, t2" 这种逗号分隔的情况
-            continue
-
-        # 处理逗号：逗号意味着可能有下一个表名，重置别名状态，保持 FROM 上下文
-        if token == ",":
-            is_join_context = False
-            # is_from_context 保持不变
-            continue
-
-    # 4. 过滤 CTE 中间表
-    # 注意：_get_cte_mid_tables 需要解析 WITH 子句，这里假设它工作正常
-    cte_tables = _get_cte_mid_tables(clean_sql)
-    source_tables -= set(cte_tables)
-
-    # 5. 构建结果
-    if source_tables or target_tables:
-        return {"source_tables": list(source_tables), "target_tables": list(target_tables)}
-
-    return None
-
-
-# ============================================================================
-# Private helper functions
-# ============================================================================
-
-
 def _trim_single_line_comment(sql: str) -> str:
     """删除单行注释"""
     result = []

{sqlh-0.3.1 → sqlh-0.3.2}/sqlh/tests/test_utils.py

@@ -3,9 +3,9 @@ from typing import Tuple
 from sqlh import utils
 
 # 读取目录或文件
-sql_path = ""
+sql_path = "/Users/dunett/codes/duperl/daas-migration/showyu_fastdata_backup_20260413"
 sql_stmt_str = utils.read_sql_from_directory(sql_path)
-sql_stmt_str = """insert into t3 SELECT /*+edede */ COUNT(*) FROM t2 JOIN [broadcast] t1 ON t1.c1 = t2.c2;"""
+# sql_stmt_str = """insert into t3 SELECT /*+edede */ COUNT(*) FROM t2 JOIN [broadcast] t1 ON t1.c1 = t2.c2;"""
 
 def test_read_sql_from_directory():
     import timeit

{sqlh-0.3.1 → sqlh-0.3.2}/sqlh/utils.py

@@ -26,7 +26,7 @@ from pathlib import Path
 from typing import List, Literal, Tuple, Union
 
 from .core.graph import DagGraph, NodeNotFoundException
-from .core.helper import get_source_target_tables_v2, split_sql, trim_comment
+from .core.helper import split_sql, trim_comment, get_source_target_tables
 
 SearchResult = Union[Tuple[List[str], DagGraph], NodeNotFoundException]
 
@@ -96,7 +96,7 @@ def __build_tables_and_graph(sql_stmt_str: str) -> Tuple[list, list, DagGraph]:
     dg = DagGraph()
 
     for sql_stmt in sql_stmt_lst:
-        table_info = get_source_target_tables_v2(sql_stmt)
+        table_info = get_source_target_tables(sql_stmt)
         if table_info:
             sources = [re.sub(r"`|\"", "", t) for t in table_info["source_tables"]]
             targets = [re.sub(r"`|\"", "", t) for t in table_info["target_tables"]]