我专注于 SQL 语句的 Where 子句,并试图撤回 Where 子句中使用的所有列。下面是一个示例 SQL:
sql_2 = """Select
PERS_ID
, STF_NO
, NAME
FROM
TEST T
WHERE T.JOIN_DT >= T.POSTING_DT'
AND T.PERS_ID LIKE '%123%'
AND T.LEAVE_DT BETWEEN CURRENT_DATE - 20 AND CURRENT_DATE - 1"""
我期待的列是 T.JOIN_DT、T.POSTING_DT、T.PERS_ID 和 T.LEAVE_DT。下面的代码能够提取除 T.PERS_ID 之外的所有内容
from __future__ import print_function
import re
import sqlparse
import numpy as np
from sqlparse.sql import IdentifierList, Identifier, Function, Where, Parenthesis, TokenList, Comparison, Operation
from sqlparse.tokens import Keyword, DML, Punctuation
sql_2 = """Select
PERS_ID
, STF_NO
, NAME
FROM
TEST T
WHERE T.JOIN_DT >= T.POSTING_DT'
AND T.PERS_ID LIKE '%123%'
AND T.LEAVE_DT BETWEEN CURRENT_DATE - 20 AND CURRENT_DATE - 1"""
parsed = sqlparse.parse(sql_2)[0]
where_columns = []
full_columns = []
for item in parsed.tokens:
if isinstance(item, Where):
for condition in item.tokens:
if isinstance(condition, Identifier):
where_columns.append(condition.get_parent_name())
where_columns.append(condition.get_real_name())
full_columns.append(where_columns)
where_columns = []
if isinstance(condition, Comparison):
for breakdown in condition.tokens:
if isinstance(breakdown, Identifier):
where_columns.append(breakdown.get_parent_name())
where_columns.append(breakdown.get_real_name())
full_columns.append(where_columns)
where_columns = []
print(full_columns)
似乎将 'AND P.PERS_ID LIKE '%123' 分组为一个令牌,我不确定如何进一步分解它以获得标识符 P.PERS_ID。如果你添加print(condition)
之后for condition in item.tokens:
你会明白我的意思。