深度学习 | 读写数据文件（定期更新，补充）

在深度学习数据预处理和结果后处理的过程中，总会和数据文件过招。所以怎样有效的读写数据文件，是一个炼丹师必备的技能，小白炼丹师打算在这篇中整理相关的代码，以便后续直接使用。

目录

• txt
• csv
• json
• h5py

读txt文件：
写txt文件：

def write_txt():
	txt_path = "..."
	with open(txt_path, 'a') as f:  #以追加的方式写入txt
		#遍历得到待存储值：x
		f.write(x + '\n')

读csv文件：

def read_csv():
	path  = "results/result_ff.csv"
	with open(path, encoding='utf-8') as f:
    	csv_reader=csv.reader(f)
   	 	for row in csv_reader:
   	 		x = row[1]

写csv文件：

def write_csv():
	path  = "results/result_ff.csv"
   	with open(path,'a+') as f:
        csv_write = csv.writer(f)
        data_row = ["path","target","result","correct","fake_frame_id"]
        csv_write.writerow(data_row) #write head of the file
        #遍历得到待存储值：x
        csv_write.writerow(x)

合并csv文件：

def csv_csv(inputpath1, inputpath2, outpath):

    csv_1=pd.read_csv(inputpath1)
    csv_2=pd.read_csv(inputpath2)
    out_csv=pd.concat([csv_1,csv_2],axis=1)   #按列合并
    out_csv.to_csv(outpath,index=False)