【Python】EXCEL转Json

在网上查阅了好几个EXCEL转Json的代码，有的是仅支持一个层级的Json（这个就不支持多层级的json），有的太过复杂看的不是很懂，所以就自己写了一个目前符合自己使用的代码。
我先讲下实现的方式。

如果遇到一个如下的json格式，我该怎么存到excel方便读取出来呢？

{
    "name": "haha",
    "value": 12,
    "floor_area": 43.5,
    "categories": [
        {
            "id": 1,
            "extra_property": "xixi",
            "renovation_type": [
                1,
                2
            ],
            "subcategories": [
                {
                    "subcategory_id": 1,
                    "subcategory_value": 2
                }
            ]
        }
    ]
}

这是个多维json，存在excel中不是很好存放，读取出来也得花点心思，毕竟你不知道后期又会有什么格式的json出现。为了应对千奇百怪的json格式，我想到如下方式

上图中特别标注的，可以划分为三个部分，以下对这三个部分进行讲解。

区域一（功能校验区域）：

 

user：登陆的用户或操作的用户

real_code：预计接口执行操作后需要返回的code值，如200、401等

real_message：预计接口执行操作后需要返回的message，如“请求成功‘、”无操作权限“等

区域二（参数类型区域）：

 

传递参数的类型无外乎这几种：int、str、float、boolean。同样我们需要告诉代码这个参数是什么类型的参数，以免传递数据的时候发生参数类型错误。在代码中我做了处理，可以识别常用的四种类型（不区分大小写）。

区域三（参数区域）：

如果是数组类型的数据，用"[]"来标记，不同的数据存储在不同的单元格中，如：

看到图的时候你会问：为什么sheet里面会有参数的名称？

 

我们可以看出“categories”是个哈希，底下还存在不同的参数。为了知道哪些参数是在“categories”下的，我们可以用一张sheet去存储“categories”参，这样我们可以通过代码方便去找到。

有可能“categories”存在多组数据，所以我们需要用"[]"来告知代码。要读取那些数据，我们可以存储行号，不同的行号可以用顿号分隔

附上代码：

# -*- coding: utf-8 -*-
# !/usr/bin/python3

import os
import re
import xlrd
import json
import unicodedata

_author_ = 'garcia'


class DataCenter:
    def __init__(self):
        # Default File Path:
        self.data_dir = os.getenv('AUTO_DATACENTER', '存放excel的文件夹地址')

        self.filename = ''
        self.path = ''
        self.sheetNames = ''

    @staticmethod
    def is_number(val):
        """判断val是否是str"""
        try:
            return True
        except ValueError:
            pass

        try:
            unicodedata.numeric(val)
            return True
        except (TypeError, ValueError):
            pass

    def keep_integer_type_from_excel(self, value):
        # Keep integer number as integer type. When reading from excel it has been changed to float type.
        if self.is_number(value) and type(value) != str and value % 1 == 0:
            return int(value)
        else:
            return value

    def change_field_type(self, table, col, val):
        # 字段类型转换
        field_type = table.cell(0, col).value
        val = self.keep_integer_type_from_excel(val)
        try:
            if val == '' or val is None:
                pass
            elif field_type.lower() == 'int':
                return int(val)
            elif field_type.lower() == 'float':
                return float(val)
            elif field_type.lower() == 'boolean':
                return int(bool(val))
            elif field_type.lower() == 'str' or field_type == '' or field_type is None:
                return str(val)
            else:
                return '字段类型错误！'
        except Exception as e:
            print(e)

    @staticmethod
    def unic(item):
        # Resolved Chinese mess code.
        try:
            item = json.dumps(item, ensure_ascii=False, encoding='UTF-8')
        except UnicodeDecodeError:
            try:
                item = json.dumps(item, ensure_ascii=False, encoding='UTF-8')
            except:
                pass
        except:
            pass

        # Convert json data string back
        item = json.loads(item, encoding='UTF-8')
        return item

    @staticmethod
    def get_sheet_names(wb):
        """
        Returns the names of all the worksheets in the current workbook.
        """
        sheet_names = wb.sheet_names()
        return sheet_names

    @staticmethod
    def __convert_to_list(val):
        """转换字符串为list"""
        value_list = re.split(',|，|、', val)
        for i in range(len(value_list)):
            value_list[i] = int(value_list[i])

        return value_list

    def get_table(self, sheet_name):
        if self.path is None:
            # Default Data Directory
            file = os.path.join(self.data_dir, self.filename)
        else:
            file = os.path.join(self.path, self.filename)

        try:
            excel_date = xlrd.open_workbook(file)
            # 得到excel的全部sheet标签值
            self.sheetNames = self.get_sheet_names(excel_date)
            my_sheet_index = self.sheetNames.index(sheet_name)
            table = excel_date.sheet_by_index(my_sheet_index)
        except Exception as e:
            print(e)

        return table

    @staticmethod
    def get_row_and_col(table):
        """获取列数、行数"""
        total_row = table.nrows
        total_col = table.ncols

        return total_row, total_col

    @staticmethod
    def get_param(table, start_col, total_col):
        param_list = []  # 获取参数
        for col in range(start_col, total_col):
            param = table.cell(1, col).value  # 获取字段名
            if param is None or param == '':
                param_list.append(param_list[-1])
            else:
                param_list.append(param)
        return param_list

    def get_child_param(self, param, row, includeEmptyCells):
        if param in self.sheetNames:
            table = self.get_table(param)
            child_total_row, child_total_col = self.get_row_and_col(table)

            child_param = self.get_param(table, 1, child_total_col)

            data_dic = {}
            count = 0
            for col in range(1, child_total_col):
                # Solve issue that get integer data from Excel file would be auto-changed to float type.
                val = self.change_field_type(table, col, table.cell(row, col).value)
                param = child_param[count]
                count += 1
                if '[]' in param:
                    if val == '' or val is None:
                        pass
                    else:
                        param = param[:param.index('[')]
                        data_dic[param] = [] if param not in data_dic.keys() else data_dic[param]
                        if param in self.sheetNames:
                            val_list = self.__convert_to_list(val)
                            for i in range(len(val_list)):
                                data_dic[param].append(
                                    self.get_child_param(param, val_list[i] - 1, includeEmptyCells))
                        else:
                            data_dic[param].append(val)
                else:
                    if param in self.sheetNames:
                        if val is not None and val != '':
                            val_list = self.__convert_to_list(val)
                            for i in range(len(val_list)):
                                data_dic[param] = self.get_child_param(param, val_list[i] - 1, includeEmptyCells)
                    elif (val == '' or val is None) and includeEmptyCells == 'False':
                        pass
                    else:
                        data_dic[param] = val
        return data_dic

    def param_to_json(self, filename, sheet_name, includeEmptyCells, path=None):
        """
        获取指定sheet中接口参数
        :param filename: 文件名
        :param sheet_name: 读取excel的sheet名称
        :param path:文件路径
        :return:
        """
        try:
            self.filename = filename
            self.path = path
            table = self.get_table(sheet_name)

            total_row, total_col = self.get_row_and_col(table)

            function_point_list = []
            check_list = []
            user_list = []
            all_data_list = []
            param_list = self.get_param(table, 4, total_col)

            for row in range(2, total_row):
                data_dic = {}
                get_check_list = []
                count = 0
                for col in range(4, total_col):
                    # Solve issue that get integer data from Excel file would be auto-changed to float type.
                    val = self.change_field_type(table, col, table.cell(row, col).value)
                    param = param_list[count]
                    count += 1
                    if '[]'in param:
                        if val == '' or val is None:
                            pass
                        else:
                            param = param[:param.index('[')]
                            data_dic[param] = [] if param not in data_dic.keys() else data_dic[param]
                            if param in self.sheetNames:
                                val_list = self.__convert_to_list(val)
                                for i in range(len(val_list)):
                                    data_dic[param].append(
                                        self.get_child_param(param, val_list[i] - 1, includeEmptyCells))
                            else:
                                data_dic[param].append(val)
                    else:
                        if param in self.sheetNames:
                            if val is not None and val != '':
                                val_list = self.__convert_to_list(val)
                                for i in range(len(val_list)):
                                    data_dic[param] = self.get_child_param(param, val_list[i] - 1, includeEmptyCells)
                            else:
                                pass
                        elif (val == '' or val is None) and includeEmptyCells == 'False':
                            pass
                        else:
                            data_dic[param] = val
                print(data_dic)
                get_check_list.append(self.keep_integer_type_from_excel(table.cell(row, 2).value))
                get_check_list.append(self.keep_integer_type_from_excel(table.cell(row, 3).value))
                check_list.append(get_check_list)
                all_data_list.append(data_dic)
                user_list.append(table.cell(row, 1).value)
                function_point_list.append(table.cell(row, 0).value)
        except Exception as e:
            print(e)

        # return all_data_list, function_point_list
        return user_list, all_data_list, function_point_list, check_list


if __name__ == '__main__':
    dc = DataCenter()
    userlist, allList, FunctionPoint, checklist = dc.param_to_json('存放数据的excel名称', 'sheet名', 'False')
    print(userlist, allList, FunctionPoint, checklist)

 说到这，我们来讲讲她的缺点：

1、如果存在多张表的嵌套，代码执行的时间比较长