python中各种文件类型的读写

本文汇总了在python中各种类型文件的读取和写入，包含文本、图像、表格、log文件、pickle文件、npy文件、npz文件等。

•  文本类型：txt文件、json文件、yaml文件
•  图像类型：使用skimage、PIL、opencv、imageio/scipy、plt库
•  表格类型：xlsx文件、csv文件
•  其他类型：log文件、pickle文件、npy文件、npz文件

文本类型

txt文件

text_name = 'test.txt'

# 文件写入
lines = ['aaa\n', 'bbb\n', 'ccc\n']
with open(text_name, 'w') as file:
    # 逐行写入
    for line in lines:
        file.write(line)

    # 一起写入
    file.writelines(lines)

# 文件读取
with open(text_name, 'r') as file:
    # 逐行读取
    while True:
        line = file.readline()
        if not line:
            break
        words = line.split()

    # 一起读取
    lines = file.readlines()
    for line in lines:
        words = line.split()

json文件

import json
text_name = 'test.json'

# 文件写入
data = {'1': 1, '2': [{'a': 1}, {'b': 2}, {'c': 3}], 3: {1: '!', '2': '?'}} # key必须是string, 否则也会自动转换为string
with open(text_name, "w") as file:
    json.dump(data, file, indent=4)

# 文件读取
with open(text_name, "r") as file:
    data = json.load(file)
print(json.dumps(data, indent=4))

yaml文件

import yaml
text_name = 'test.yaml'

# 文件写入
data = {'1': 1, '2': [{'a': 1}, {'b': 2}, {'c': 3}], 3: {1: '!', '2': '?'}}
with open(text_name, "w") as file:
    yaml.dump(data, file, indent=4)

# 文件读取
with open(text_name, "r") as file:
    data = yaml.load(file, Loader=yaml.FullLoader)
print(yaml.dump(data, indent=4))

图像类型

使用skimage

from skimage import io
image_name = "test.jpg"

# 文件读取
img = io.imread(image_name)
print(type(img), img.dtype, img.shape) # numpy类型, [height, width, channel]

# 文件写入
io.imsave(image_name, img)

使用PIL

from PIL import Image
import numpy as np
image_name = "test.jpg"

# 文件读取
img = Image.open(image_name)
print(type(img), img.size) # PIL类型, [height, width]
img = np.array(img)
print(type(img), img.dtype, img.shape) # numpy类型, [height, width, channel]

# 文件写入
img = Image.fromarray(img)
img.save(image_name)

使用opencv

import cv2
image_name = "test.jpg"

# 文件读取
img = cv2.imread(image_name)
print(type(img), img.dtype, img.shape) # numpy类型, [height, width, channel]

# 文件写入
cv2.imwrite(image_name, img)

使用imageio/scipy

import imageio
from scipy import misc # 新版本scipy中接口已被移除，建议使用imageio替代
image_name = "test.jpg"

# 文件读取
img = imageio.imread(image_name)
print(type(img), img.dtype, img.shape) # imageio类型, [height, width, channel]

# 文件写入
imageio.imwrite(image_name, img)

使用plt

import matplotlib.pyplot as plt
image_name = "test.jpg"

# 文件读取
img = plt.imread(image_name)
print(type(img), img.dtype, img.shape) # numpy类型, [height, width, channel]

# 文件写入
plt.imsave(image_name, img)

表格类型

xlsx文件

import openpyxl
table_name = "test.xlsx"

# 文件写入
workbook = openpyxl.Workbook()
sheet = workbook.active
sheet.title = 'Sheet1'
sheet.cell(1, 1, "TEST") # 下标从[1, 1]开始

line1 = ['aaa', 1]
line2 = ['bbb', 2]
sheet.append(line1)
sheet.append(line2)

workbook.save(table_name)

# 文件读取
workbook = openpyxl.load_workbook(table_name)
sheet = workbook.get_sheet_by_name('Sheet1') # or sheet = workbook.active
print(sheet.title, sheet.max_row, sheet.max_column)
cell = sheet.cell(1, 1)
print(cell.value)

csv文件

import csv
table_name = 'test.csv'

# 文件写入
line1 = ['aaa', 1]
line2 = ['bbb', 2]
with open(table_name, 'w', newline='') as file:
    csv_write = csv.writer(file, dialect='excel')
    csv_write.writerow(line1)
    csv_write.writerow(line2)

# 文件读取
with open(table_name, 'r', newline='') as file:
    reader = csv.reader(file)
    for row in reader:
        print(row)

其他类型

log文件

import logging

logger = logging.getLogger()
logger.setLevel(level=logging.DEBUG)

# 输出到文件
log_name = 'test.log'
filehandler = logging.FileHandler(log_name, mode='w')
filehandler.setLevel(level=logging.INFO)
formatter = logging.Formatter("%(asctime)s - %(filename)s [line:%(lineno)d] - %(levelname)s: %(message)s")
filehandler.setFormatter(formatter)
logger.addHandler(filehandler)

# 输出到屏幕
consolehandler = logging.StreamHandler()
consolehandler.setLevel(level=logging.DEBUG)
logger.addHandler(consolehandler)

logger.info("INFO")
logger.debug("DEBUG")

pickle文件

import pickle
file_name = "test.pickle"

# 写入文件
string = [a*'a' for a in range(10)]
label = [i for i in range(10)]
with open(file_name, 'wb') as f:
    pickle.dump([string, label], f, protocol=2)

# 读取文件
with open(file_name, 'rb') as f:
    string, label = pickle.load(f)
    print(string, label)

npy文件

npy文件用于保存单个numpy类型的数组

import numpy as np
file_name = "test.npy"

# 写入文件
data = np.array([[1, 2], [3, 4]])
np.save(file_name, data)

# 读取文件
data = np.load(file_name)
print(data)

npz文件

npz文件用于同时保存多个numpy类型的数组

import numpy as np
file_name = "test.npz"

# 写入文件
data = {"data_1": np.array([[1, 2], [3, 4]]), "data_2": np.array([[1, 2], [3, 4]]) * 2}
np.savez(file_name, **data)

# 读取文件
data = np.load(file_name)
keys = list(data)
print(keys, data['data_1'], data['data_2'])