[英]Python - Most efficient to overwrite a specific row in a CSV file
Given the following csv file : 给定以下csv文件:
01;blue;brown;black
02;glass;rock;paper
03;pigeon;squirel;shark
My goal is to replace the (unique) line containing '02' in the 1st posisition. 我的目标是在第一个病房中替换包含“ 02”的(唯一)行。
I wrote this piece of code: 我写了这段代码:
with open("csv", 'r+', newline='', encoding='utf-8') as csvfile, open('csvout', 'w', newline='', encoding='utf-8') as out:
reader = csv.reader(csvfile, delimiter=';')
writer = csv.writer(out, delimiter=';')
for row in reader:
if row[0] != '02':
writer.writerow(row)
else:
writer.writerow(['02', 'A', 'B', 'C'])
But re-writing the whole CSV in an other doesn't seem to be the most efficient way to proceed, especially for large files: 但是,以另一种方式重写整个CSV似乎并不是最有效的处理方式,尤其是对于大文件:
I wrote a second piece of code who seems to answer to these two problems : 我写了第二段代码,似乎可以回答这两个问题:
with open("csv", 'r+', newline='', encoding='utf-8') as csvfile:
content = csvfile.readlines()
for index, row in enumerate(content):
row = row.split(';')
if row[2] == 'rock':
tochange = index
break
content.pop(tochange)
content.insert(tochange, '02;A;B;C\n')
content = "".join(content)
csvfile.seek(0)
csvfile.truncate(0) # Erase content
csvfile.write(content)
Do you agree that the second solution is more efficient ? 您是否同意第二种解决方案更有效? Do you have any improvement, or better way to proceed ?
您有什么改进或更好的方法吗?
EDIT : The number of character in the line can vary. 编辑:该行中的字符数可以变化。
EDIT 2 : I'm apparently obliged to read and rewrite everything, if I don't want to use padding. 编辑2 :如果我不想使用填充,显然我必须阅读和重写所有内容。 A possible solution would be a database-like solution, I will consider it for the future.
可能的解决方案是类似数据库的解决方案,我将在以后考虑。
If I had to choose between those 2 solutions, which one would be the best performance-wise ? 如果我必须在这两种解决方案之间进行选择,哪种才是最佳性能选择?
As the caracter in the line may vary, I either have to read/write the whole file or; 由于该行中的角色可能有所不同,我要么必须读/写整个文件,要么; as @tobias_k said, use seek() to come back to the begining of the line and:
正如@tobias_k所说,使用seek()返回到行的开头 ,并:
I want to avoid using padding so I used time.perf_counter() to measure exec time of both codes, and the second solution appears to be (almost 2*) faster (CSV of 10 000 lines, match at the 6 000th). 我想避免使用填充,因此我使用time.perf_counter()来测量两个代码的执行时间, 第二种解决方案似乎快了 ( 几乎2 *) (10,000行的CSV,在第6000行匹配)。
One alternative would be to migrate to a relational database . 一种选择是迁移到关系数据库 。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.