nsq_convert2txt/file_processing/file_xlsx.py

"""
coding=uft-8
解.xlsx文件

单元格ctype的五种类型
ctype : 0 empty,1 string, 2 number, 3 date, 4 boolean, 5 error
"""

from loguru import logger
from docs.config import error_number
import pandas as pd


def read_xlsx(file_path: str):
    """
    返回拼接好的字符串
    :param file_path:
    :return:
    """
    try:
        excel_data = pd.read_excel(file_path, keep_default_na=False, sheet_name=None)
        tables = []
        for sheet_name, table_2d in excel_data.items():
            string = '<table>'
            for row in table_2d.values:
                string += '<tr>'
                for clown in row:
                    clown = str(clown)
                    string += '<td>' + clown.strip('\n').strip('\t') + '</td>'
                string += '</tr>'
            string_end = string + '</table>'
            tables.append(string_end)
        content = "\n".join(tables)
        return content, error_number["成功"]
    except Exception as e:
        logger.error('xlsx文件解析出错')
        logger.error(e)
        return '', error_number["解析错误"]


if __name__ == '__main__':
    file_path = "0.xls"
    string = read_xlsx(file_path)
    print(string)