20240410

文件管理1.ipynb

@@ -615,6 +615,87 @@
     "print('ok')  "
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "75aa9852-64aa-4ede-b9a0-00cd4f672527",
+   "metadata": {},
+   "source": [
+    "## word文档读取表格"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2d01d0ae-ab9a-4798-9723-51e4fd2bb112",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "import docx\n",
+    "import json\n",
+    "doc = docx.Document(r\"data/中国石化北海炼化有限责任公司2023年健康体检团体报告.docx\")\n",
+    "for i, table in enumerate(doc.tables):\n",
+    "        print(f\"Table {i}:\")\n",
+    "        for row in table.rows:\n",
+    "            print(type(row))\n",
+    "            for cell in row.cells:\n",
+    "                if cell.text =='序号':\n",
+    "                    print(cell.text, end=\" | \")\n",
+    "            print()  # 每一行结束后换行\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "c3ccfa25-f31c-46d4-bfd0-903832cc620d",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2024-04-10T11:08:13.063231Z",
+     "iopub.status.busy": "2024-04-10T11:08:13.063012Z",
+     "iopub.status.idle": "2024-04-10T11:08:25.173599Z",
+     "shell.execute_reply": "2024-04-10T11:08:25.173048Z",
+     "shell.execute_reply.started": "2024-04-10T11:08:13.063213Z"
+    },
+    "tags": []
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "ok!\n"
+     ]
+    }
+   ],
+   "source": [
+    "import docx\n",
+    "import json\n",
+    "import openpyxl\n",
+    "doc = docx.Document(r\"data/中国石化北海炼化有限责任公司2023年职业体检团体报告.docx\")\n",
+    "list2 = []\n",
+    "for i, table in enumerate(doc.tables):\n",
+    "    \n",
+    "    if table.cell(0, 0).text == '序号':\n",
+    "        #print(f\"Table {i}:\")\n",
+    "        \n",
+    "        for row in table.rows:\n",
+    "            list1= []\n",
+    "            for cell in row.cells:\n",
+    "                list1.append(cell.text)\n",
+    "            if list1[0] !='序号':\n",
+    "                list2.append(list1)\n",
+    "filename = '北海体测报告情况表2024.xlsx'\n",
+    "wb = openpyxl.Workbook()\n",
+    "sheet = wb.active\n",
+    "\n",
+    "for row in list2:\n",
+    "    sheet.append(row)\n",
+    "    \n",
+    "wb.save(filename)\n",
+    "print('ok!')"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -735,27 +816,12 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 40,
+   "execution_count": null,
    "id": "d05d85d2-5229-446e-bfff-1f0feb8b7f09",
    "metadata": {
-    "execution": {
-     "iopub.execute_input": "2023-04-10T10:04:32.991821Z",
-     "iopub.status.busy": "2023-04-10T10:04:32.991387Z",
-     "iopub.status.idle": "2023-04-10T10:04:34.154437Z",
-     "shell.execute_reply": "2023-04-10T10:04:34.153707Z",
-     "shell.execute_reply.started": "2023-04-10T10:04:32.991783Z"
-    },
     "tags": []
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "13\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "import openpyxl\n",
     "from docxtpl import DocxTemplate,InlineImage\n",
@@ -1175,7 +1241,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.6"
+   "version": "3.10.12"
   }
  },
  "nbformat": 4,