python 提取 Excel 特定列 - 帶界面

需求 : 一個 Excel 有多個列, 需要提取部分的列. 生成一個新的 Excel

python 依賴 tkinter, pandas

 

# -*- coding: UTF-8 -*-
import tkinter as tk
from tkinter import filedialog, RIGHT, Scrollbar, Y, LEFT
from os import walk
import pandas as pd
import random


class Application(tk.Frame):

    # 記錄 excel 表頭
    excel_columns = []
    # 記錄對應的 excel 列的狀態
    excel_columns_checkbutton_state = []
    directory_path = ''
    def __init__(self, master=None):

        super().__init__(master)
        self.export_button = tk.Button(self)
        self.quit = tk.Button(self, text="退出", fg="red", command=self.master.destroy)
        self.file_dialog = tk.Button(self)
        self.master = master
        self.pack()
        self.create_widgets()

    def create_widgets(self):
        # 選擇日誌文件的按鈕
        self.file_dialog["text"] = "選擇日誌文件"
        self.file_dialog["command"] = self.create_file_dialog
        self.file_dialog.pack(side="left")

        # 導出 Excel 的列
        self.export_button["text"] = "導出選中列"
        self.export_button["command"] = self.export_excel_column
        self.export_button.pack(side=LEFT)

        # 退出按鈕
        self.quit.pack(side="left")

    def export_excel_column(self):
        print(list(map((lambda var: var.get()), self.excel_columns_checkbutton_state)))
        result_data = pd.DataFrame()  # 建立一個空的dataframe
        target_export_columns = []
        for i in range(len(self.excel_columns_checkbutton_state)):
            if self.excel_columns_checkbutton_state[i].get() == 1:
                target_export_columns.append(self.excel_columns[i])

        for path, dirs, files in walk(self.directory_path, topdown=False):  # 這裏我們的數據都存儲在文件夾下
            print(files)

        print("要導出的列是: ", target_export_columns)
        print(self.excel_columns)
        for i in range(len(files)):
            # 讀取每個 excel 文件中的數據
            target_data = pd.read_excel(self.directory_path + '/%s' % files[i], usecols=target_export_columns)
            print("target_data 的類型是: ", type(target_data))
            col_mean = target_data[target_export_columns].mean()

            # 將每個excel中的數據存儲到之前建好的空的 data frame 中
            result_data = result_data.append(col_mean, ignore_index=True)

        writer = pd.ExcelWriter(self.directory_path + '/output.xlsx')
        result_data.to_excel(writer, 'AllData')  # 這裏“AllData”是sheet的名字
        writer.save()  # 執行完這一步之後，合併後的表格就保存

    def create_file_dialog(self):
        file_path = filedialog.askdirectory()
        self.directory_path = file_path
        for directory_path, dirs, files in walk(file_path, topdown=False):  # 這裏我們的數據都存儲在文件夾下
            print(files)
        print("files 類型是:", type(files))
        # 得到 Excel 表頭
        first_data_frame = pd.read_excel(file_path + '/%s' % files[0])
        print(first_data_frame.columns.values)
        print(type(first_data_frame.columns.values))

        # 生成 Excel 表頭的 checkbutton
        excel_columns_count = -1
        rows_count = 0
        columns_count = 0
        print(len(first_data_frame.columns.values))
        for col in first_data_frame.columns.values:
            excel_columns_count = excel_columns_count + 1
            print(col)
            # 記錄 Excel 表頭的列名
            self.excel_columns.append(col)
            y = tk.IntVar()
            self.excel_columns_checkbutton_state.append(y)
            ct = [random.randrange(256) for x in range(3)]
            brightness = int(round(0.299 * ct[0] + 0.587 * ct[1] + 0.114 * ct[2]))
            ct_hex = "%02x%02x%02x" % tuple(ct)
            bg_colour = '#' + "".join(ct_hex)

            b = tk.Checkbutton(self.master, text=col, fg='White' if brightness < 120 else 'Black', bg=bg_colour, variable=y)
            if excel_columns_count % 15 == 0:
                columns_count = columns_count + 1
                rows_count = 0

            rows_count = rows_count + 1
            b.place(x=250 * columns_count, y=rows_count * 70)


if __name__ == '__main__':
    root = tk.Tk()
    # scrollbar = Scrollbar(root)
    # scrollbar.pack(side=RIGHT, fill=Y)
    root.geometry("1500x900+200+50")
    app = Application(master=root)
    app.mainloop()