mdbq 2.5.6__py3-none-any.whl → 2.5.8__py3-none-any.whl

mdbq/clean/clean_upload.py

@@ -7,7 +7,6 @@ import zipfile
 from pyzipper import PyZipFile
 import os
 import platform
-import pathlib
 import json
 from mdbq.mongo import mongo
 from mdbq.mysql import mysql
@@ -711,7 +710,7 @@ class DataClean:
             _date = re.findall(r'(\d{4}-\d{2})-\d{2}', str(_name))
             if _date:
                 _date = _date[0]
-                t2 = pathlib.Path(t2, _date)  # 添加 年月分类
+                t2 = os.path.join(t2, _date)  # 添加 年月分类
                 if not os.path.exists(t2):
                     os.makedirs(t2, exist_ok=True)
         old_file = os.path.join(t2, _name)  # 检查目标位置是否已经存在该文件
@@ -1213,6 +1212,49 @@ def main(service_databases=None):
     )
 
 
+def test():
+    # main_key = '单元报表'
+    path = f'/Users/xigua/数据中心/原始文件2/京东报表/JD流量来源13414124124'
+    for root, dirs, files in os.walk(path, topdown=False):
+        for name in files:
+            if '~$' in name or '.DS' in name or '.localized' in name or '.jpg' in name or '.png' in name:
+                continue
+            if 'py_xg' in name:
+                continue
+            if 'TM_旧表_字段' in root:
+                continue
+
+            if name.endswith('.csv'):
+                print(name)
+                df = pd.read_csv(os.path.join(root, name), encoding='utf-8_sig', header=0, na_filter=False)
+                # if '店铺名称' not in df.columns.tolist():
+                #     df.insert(loc=1, column='店铺名称', value='京东箱包旗舰店')
+                for item in df.columns.tolist():
+                    if '同比' in item or '环比' in item or '时间' in item:
+                        df.pop(item)
+                date_all = re.findall(r'_(\d{4}-\d{2}-\d{2})_\d{4}-\d{2}-\d{2}', name)[0]
+
+                date = re.findall(r'_(\d{4}-\d{2})-\d{2}', name)[0]
+
+                new_path = f'/Users/xigua/数据中心/原始文件3/京东报表/店铺来源_三级来源/{date}'
+                # new_path = os.path.join(new_path, date)  # 添加 年月分类
+                if not os.path.exists(new_path):
+                    os.makedirs(new_path, exist_ok=True)
+                # print(date_all)
+
+                new_name = f'py_xg_京东商智_店铺来源_三级来源_{date_all}.xlsx'
+                # print(os.path.join(new_path, new_name))
+                # breakpoint()
+                # df.to_csv(os.path.join(new_path, new_name), encoding='utf-8_sig', index=False, header=True)
+                try:
+                    df.to_excel(os.path.join(new_path, new_name),
+                                index=False, header=True, engine='openpyxl', freeze_panes=(1, 0))
+                except Exception as e:
+                    print(e)
+
+
+
+
 if __name__ == '__main__':
     main(
         service_databases = [
@@ -1230,3 +1272,6 @@ if __name__ == '__main__':
     # )
     # c.sp_scene_clean(is_except=['except'])  # 商品素材
     # c.move_tg_tm(is_except=['临时文件', ])  # 天猫，移到文件到原始文件夹
+
+
+    # test()

mdbq/dataframe/converter.py

@@ -47,7 +47,7 @@ class DataFrameConverter(object):
             try:
                 # 百分比在某些数据库中不兼容, 转换百分比为小数, # 转百分比的列不能含有中文或特殊字符
                 df[col] = df[col].apply(
-                    lambda x: float(float((str(x).rstrip("%"))) / 100) if re.findall(r'^\d+%', str(x)) else x)
+                    lambda x: float(float((str(x).rstrip("%"))) / 100) if re.findall(r'^\d+\.?\d*%', str(x)) else x)
             except Exception as e:
                 print(f'留意错误信息: 位于列 -> {col} -> {e}')
 
@@ -93,6 +93,6 @@ if __name__ == '__main__':
     # df = converter.convert_df_cols(df)
     # print(df['a'].dtype)
     # print(df)
-    pattern = 'dfa_dfawr__'
-    pattern = re.sub(r'_+$', '', pattern)
+    pattern = '1540%'
+    pattern = re.findall(r'^\d+\.?\d*%', pattern)
     print(pattern)

{mdbq-2.5.6.dist-info → mdbq-2.5.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: mdbq
-Version: 2.5.6
+Version: 2.5.8
 Home-page: https://pypi.org/project/mdbq
 Author: xigua, 
 Author-email: 2587125111@qq.com

{mdbq-2.5.6.dist-info → mdbq-2.5.8.dist-info}/RECORD

@@ -9,7 +9,7 @@ mdbq/aggregation/query_data.py,sha256=WKe42Xq1Gi-ELuIT0k2jh3X4-R7heb0ub3Mj3yuCRA
 mdbq/bdup/__init__.py,sha256=AkhsGk81SkG1c8FqDH5tRq-8MZmFobVbN60DTyukYTY,28
 mdbq/bdup/bdup.py,sha256=LAV0TgnQpc-LB-YuJthxb0U42_VkPidzQzAagan46lU,4234
 mdbq/clean/__init__.py,sha256=A1d6x3L27j4NtLgiFV5TANwEkLuaDfPHDQNrPBbNWtU,41
-mdbq/clean/clean_upload.py,sha256=q_3kjiE0sU6uV13TW9rVuPmbO01itYhkC4gTnz_nZ5o,64455
+mdbq/clean/clean_upload.py,sha256=mkJvqW0ewhYelNsAKw_6ajV8eggzRFVgTYaPPYFt3Ak,66478
 mdbq/clean/data_clean.py,sha256=ucfslhqXVZoH2QaXHSAWDky0GhIvH9f4GeNaHg4SrFE,104790
 mdbq/company/__init__.py,sha256=qz8F_GsP_pMB5PblgJAUAMjasuZbOEp3qQOCB39E8f0,21
 mdbq/company/copysh.py,sha256=NvlXCBZBcO2GIT5nLRYYqhOyHWM1-1RE7DHvgbj6jmQ,19723
@@ -20,7 +20,7 @@ mdbq/config/products.py,sha256=hN9UMkM6j76HYMulTYdtr3mOhh9QdpvvrLH14a_mbFY,5980
 mdbq/config/set_support.py,sha256=xkZCX6y9Bq1ppBpJAofld4B2YtchA7fl0eT3dx3CrSI,777
 mdbq/config/update_conf.py,sha256=taL3ZqKgiVWwUrDFuaYhim9a72Hm4BHRhhDscJTziR8,4535
 mdbq/dataframe/__init__.py,sha256=2HtCN8AdRj53teXDqzysC1h8aPL-mMFy561ESmhehGQ,22
-mdbq/dataframe/converter.py,sha256=u7rQbIsgVZWOIybJaknduf7wViBdBkyU8mwUo24mDt0,4304
+mdbq/dataframe/converter.py,sha256=SJLZ96f6QBxnPcaaKDi3UOLNk0b7O6aWLAyDOMiSy80,4312
 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
@@ -42,7 +42,7 @@ mdbq/req_post/__init__.py,sha256=jso1oHcy6cJEfa7udS_9uO5X6kZLoPBF8l3wCYmr5dM,18
 mdbq/req_post/req_tb.py,sha256=PexWSCPJNM6Tv0ol4lAWIhlOwsAr_frnjtcdSHCFiek,36179
 mdbq/spider/__init__.py,sha256=RBMFXGy_jd1HXZhngB2T2XTvJqki8P_Fr-pBcwijnew,18
 mdbq/spider/aikucun.py,sha256=KdihSB3q44jsXUQAldfWRVfCSrEw2MNbM-_BhP_29g4,14448
-mdbq-2.5.6.dist-info/METADATA,sha256=IIsrPFdnbkCfH4ziUSl_U05g3Cvc9vwyySZjLXN6SVU,243
-mdbq-2.5.6.dist-info/WHEEL,sha256=cpQTJ5IWu9CdaPViMhC9YzF8gZuS5-vlfoFihTBC86A,91
-mdbq-2.5.6.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
-mdbq-2.5.6.dist-info/RECORD,,
+mdbq-2.5.8.dist-info/METADATA,sha256=nu5MMMIj05LwXRKsmJwUs1e95SopXUec2w3znVPII7s,243
+mdbq-2.5.8.dist-info/WHEEL,sha256=cpQTJ5IWu9CdaPViMhC9YzF8gZuS5-vlfoFihTBC86A,91
+mdbq-2.5.8.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
+mdbq-2.5.8.dist-info/RECORD,,