mdbq 0.1.4__py3-none-any.whl → 0.1.6__py3-none-any.whl

mdbq/dataframe/converter.py

@@ -24,6 +24,7 @@ class DataFrameConverter(object):
         df.replace(to_replace=['="'], value='', regex=True, inplace=True)  # ="和"不可以放在一起清洗, 因为有: id=86785565
         df.replace(to_replace=['"'], value='', regex=True, inplace=True)
         cols = df.columns.tolist()
+
         for col in cols:
             # df[col] = df[col].apply(lambda x: re.sub('[="]', '', str(x)) if '="' in str(x) else x)
             # 百分比在某些数据库中不兼容, 转换百分比为小数
@@ -31,9 +32,14 @@ class DataFrameConverter(object):
             # 尝试转换合适的数据类型
             if df[col].dtype == 'object':
                 try:
-                    df[col] = df[col].astype(int)  # 尝试转换 int
+                    # df[col] = df[col].astype(int)  # 尝试转换 int
+                    df[col] = df[col].apply(lambda x: int(x) if '_' not in str(x) else x)
                 except:
-                    df[col] = df[col].astype('float64', errors='ignore')    # 尝试转换 float, 报错则忽略
+                    # df[col] = df[col].astype('float64', errors='ignore')    # 尝试转换 float, 报错则忽略
+                    try:
+                        df[col] = df[col].apply(lambda x: float(x) if '_' not in str(x) else x)
+                    except:
+                        pass
             if df[col].dtype == 'float':  # 对于小数类型, 保留 6 位小数
                 df[col] = df[col].apply(lambda x: round(float(x), 6) if x != 0 else x)
             # 清理列名, 在 mysql 里面列名不能含有某些特殊字符

mdbq/mysql/s_query.py

@@ -86,6 +86,33 @@ class QueryDatas:
             print(f'database: {db_name}, table: {tabel_name} 查询的数据为空')
         return df
 
+    def columns_to_list(self, db_name, tabel_name,  columns_name) -> list:
+        """
+        获取数据表的指定列, 返回列表
+        [{'视频bv号': 'BV1Dm4y1S7BU', '下载进度': 1}, {'视频bv号': 'BV1ov411c7US', '下载进度': 1}]
+        """
+        if self.check_infos(db_name, tabel_name) == False:  # 检查传入的数据库和数据表是否存在
+            return []
+
+        self.config.update({'database': db_name})
+        connection = pymysql.connect(**self.config)  # 重新连接数据库
+        with connection.cursor() as cursor:
+            # 3. 获取数据表的所有列信息
+            sql = 'SELECT COLUMN_NAME FROM information_schema.columns WHERE table_schema = %s AND table_name = %s'
+            cursor.execute(sql, (db_name, {tabel_name}))
+            columns = cursor.fetchall()
+            cols_exist = [col['COLUMN_NAME'] for col in columns]  # 数据表的所有列, 返回 list
+            columns_name = [item for item in columns_name if item in cols_exist]
+            if len(columns_name) == 0:
+                return []
+            columns_in = ', '.join(columns_name)
+            sql = (f"SELECT {columns_in} FROM {db_name}.{tabel_name} ")
+            cursor.execute(sql)
+            column_values = cursor.fetchall()  # 返回指定列，结果是[dict, dict, dict, ...]
+            # column_values = [item[column_name] for item in column_values]  # 提取字典的值, 组成列表
+        connection.close()
+        return column_values
+
     def check_infos(self, db_name, tabel_name) -> bool:
         """ 检查数据库、数据表是否存在 """
         connection = pymysql.connect(**self.config)  # 连接数据库
@@ -119,5 +146,9 @@ class QueryDatas:
 
 
 if __name__ == '__main__':
-    username, password, host, port = get_myconf.select_config_values(target_service='company', database='mysql')
+    username, password, host, port = get_myconf.select_config_values(target_service='home_lx', database='mysql')
     print(username, password, host, port)
+
+    q = QueryDatas(username, password, host, port)
+    res = q.columns_to_list(db_name='视频数据', tabel_name='bilibili视频', columns_name=['视频bv号', '下载进度'])
+    print(res)

{mdbq-0.1.4.dist-info → mdbq-0.1.6.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: mdbq
-Version: 0.1.4
+Version: 0.1.6
 Home-page: https://pypi.org/project/mdbsql
 Author: xigua, 
 Author-email: 2587125111@qq.com

{mdbq-0.1.4.dist-info → mdbq-0.1.6.dist-info}/RECORD

@@ -13,14 +13,14 @@ mdbq/config/__init__.py,sha256=jso1oHcy6cJEfa7udS_9uO5X6kZLoPBF8l3wCYmr5dM,18
 mdbq/config/get_myconf.py,sha256=9v3xebfcS1tptxpvk3_tGxfXjAehGVCveYe4iRUzLQQ,6372
 mdbq/config/update_conf.py,sha256=YjGjjRchu5BcrmLJkoLjHEF2TbGOmsgCWX4LroXOYWQ,3455
 mdbq/dataframe/__init__.py,sha256=2HtCN8AdRj53teXDqzysC1h8aPL-mMFy561ESmhehGQ,22
-mdbq/dataframe/converter.py,sha256=5hrGx-lPVwYLuyZNOHf6K7O9_AAKZ7mCp0MHxlxSCnk,2816
+mdbq/dataframe/converter.py,sha256=h_BDc6oNmMCVFOUzZJq4nXNGDyJFJyycpWqlyrv7U04,3089
 mdbq/log/__init__.py,sha256=Mpbrav0s0ifLL7lVDAuePEi1hJKiSHhxcv1byBKDl5E,15
 mdbq/log/mylogger.py,sha256=oaT7Bp-Hb9jZt52seP3ISUuxVcI19s4UiqTeouScBO0,3258
 mdbq/mongo/__init__.py,sha256=SILt7xMtQIQl_m-ik9WLtJSXIVf424iYgCfE_tnQFbw,13
 mdbq/mongo/mongo.py,sha256=q0B4wXDSTtXg_vMN7MPh6zdxl6tT68tM74LmdVNQQek,31892
 mdbq/mysql/__init__.py,sha256=A_DPJyAoEvTSFojiI2e94zP0FKtCkkwKP1kYUCSyQzo,11
 mdbq/mysql/mysql.py,sha256=w_FDE4ulOnHdfIjjXB1Ff_cnXyx2PuBAQWGjTYyVm3U,30629
-mdbq/mysql/s_query.py,sha256=pzxRGBRP3Ku4oVLqfpMd1VWWct3YlxxnVavtKa9kgSM,5302
+mdbq/mysql/s_query.py,sha256=mNrdyMeiQ5wQVD_9IDTrk6jrMtszQftYQg3AwNWhROc,6997
 mdbq/mysql/year_month_day.py,sha256=VgewoE2pJxK7ErjfviL_SMTN77ki8GVbTUcao3vFUCE,1523
 mdbq/other/__init__.py,sha256=jso1oHcy6cJEfa7udS_9uO5X6kZLoPBF8l3wCYmr5dM,18
 mdbq/other/porxy.py,sha256=UHfgEyXugogvXgsG68a7QouUCKaohTKKkI4RN-kYSdQ,4961
@@ -30,7 +30,7 @@ mdbq/pbix/__init__.py,sha256=Trtfaynu9RjoTyLLYBN2xdRxTvm_zhCniUkVTAYwcjo,24
 mdbq/pbix/pbix_refresh.py,sha256=JUjKW3bNEyoMVfVfo77UhguvS5AWkixvVhDbw4_MHco,2396
 mdbq/pbix/refresh_all.py,sha256=wulHs4rivf4Mi0Pii2QR5Nk9-TBcvSwnCB_WH9QULKE,5939
 mdbq/spider/__init__.py,sha256=RBMFXGy_jd1HXZhngB2T2XTvJqki8P_Fr-pBcwijnew,18
-mdbq-0.1.4.dist-info/METADATA,sha256=36-EtjU3OviSuiID5CH5XrLaHodQW3vIYXwtv6d2acI,245
-mdbq-0.1.4.dist-info/WHEEL,sha256=eOLhNAGa2EW3wWl_TU484h7q1UNgy0JXjjoqKoxAAQc,92
-mdbq-0.1.4.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
-mdbq-0.1.4.dist-info/RECORD,,
+mdbq-0.1.6.dist-info/METADATA,sha256=a5jG4lDS-9TA8xU5IInf--XK_X0P84L9FOZQY-ysGHU,245
+mdbq-0.1.6.dist-info/WHEEL,sha256=eOLhNAGa2EW3wWl_TU484h7q1UNgy0JXjjoqKoxAAQc,92
+mdbq-0.1.6.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
+mdbq-0.1.6.dist-info/RECORD,,