import pandas as pd from collections import Counter # 金蝶正查 kd = pd.read_excel('物料清单正查_1.xlsx', sheet_name='树形单据体') kd = kd[kd['子项物料编码'].notna()].copy() kd = kd[kd['BOM层级'].astype(str).str.strip() != '0'] kd['code'] = kd['子项物料编码'].astype(str).str.strip() kd['name'] = kd['物料名称'].astype(str).str.strip() # 金蝶父级:由 BOM层级 推导(1.1.2 的父级是 1.1) def kd_parent(lv): parts = str(lv).split('.') return '.'.join(parts[:-1]) if len(parts) > 1 else '' kd['lv'] = kd['BOM层级'].astype(str).str.strip() kd['parent_lv'] = kd['lv'].apply(kd_parent) # 层级路径 -> 编码 映射(含根 0) lv2code = {'0': 'CG0S175E0000120'} for _, r in kd.iterrows(): lv2code[r['lv']] = r['code'] kd['parent'] = kd['parent_lv'].map(lv2code) db = pd.read_csv('db_bom2.csv', dtype={'ChildMaterialId': str, 'ParentMaterialId': str}) db['code'] = db['ChildMaterialId'].astype(str).str.strip() db['parent'] = db['ParentMaterialId'].astype(str).str.strip() # 根物料在数据库里父级可能为空,统一填根编码 db['parent'] = db['parent'].replace({'': 'CG0S175E0000120', 'nan': 'CG0S175E0000120', 'None': 'CG0S175E0000120'}) kd['parent'] = kd['parent'].fillna('CG0S175E0000120').astype(str) print(f"金蝶子项: {len(kd)}, 数据库: {len(db)}") # 按 (父级, 子件) 对比 —— 用户要求:哪些组件下子件个数不一样 kd_pc = Counter(zip(kd['parent'], kd['code'])) db_pc = Counter(zip(db['parent'], db['code'])) all_parents = sorted(set(kd['parent']) | set(db['parent'])) print("\n=== 各组件(父级)子件行数对比 ===") diff_rows = [] for p in all_parents: k = sum(v for (pp, c), v in kd_pc.items() if pp == p) d = sum(v for (pp, c), v in db_pc.items() if pp == p) if k != d: diff_rows.append((p, k, d)) print(f" 父级 {p}: 金蝶 {k} 行 vs 数据库 {d} 行") if not diff_rows: print(" (无差异)") # 每个有差异的父级,列出具体多/少的子件 print("\n=== 差异明细(按父级)===") name_map = {} for _, r in kd.iterrows(): name_map.setdefault(r['code'], r['name']) for _, r in db.iterrows(): name_map.setdefault(r['code'], r.get('MaterialName', '')) for p, k, d in diff_rows: kp = {c: v for (pp, c), v in kd_pc.items() if pp == p} dp = {c: v for (pp, c), v in db_pc.items() if pp == p} print(f"\n父级 [{p}] {name_map.get(p,'')}: 金蝶{k}行 vs 数据库{d}行") only_kd = set(kp) - set(dp) only_db = set(dp) - set(kp) both_diff = {c for c in set(kp) & set(dp) if kp[c] != dp[c]} for c in sorted(only_kd): print(f" 仅金蝶有: {c} x{kp[c]} ({name_map.get(c,'')})") for c in sorted(only_db): print(f" 仅数据库有: {c} x{dp[c]} ({name_map.get(c,'')})") for c in sorted(both_diff): print(f" 次数不同: {c} 金蝶x{kp[c]} vs 数据库x{dp[c]} ({name_map.get(c,'')})")