在文本文件中存储Python实例属性的更好方法是什么？

Question

I need to store a relatively large list of changeable instance variables in a text file. This is a breakdown of my current method for reading and writing:

class Car():
    def __init__(self, name, size, color):
        self.name = name
        self.size = size
        self.color = color

aa = Car('truck','big','red')
bb = Car('sedan','small','blue')
cc = Car('bus','big','yellow')

with open('test.csv', 'ab') as ff:
    ww = csv.writer(ff)
    for i in [aa, bb, cc]:
        ww.writerow([i.name, i.size, i.color])

carList = []
with open('test.csv', 'rb') as ff:
    rr = csv.reader(ff)
    for i in rr:
        carList.append(Car(i[0], i[1], i[2]) )

print carList[1].color

This has a few problems. The number and order of columns are effectively hardcoded in. I am not sure if this can be easily avoided, but I feel like it should be. Modifying an instance in the middle of the file would, I believe, require iterating through all lines and rewriting them while checking 'name's (which are unique unlike the other variables) for a line to write differently. Also I can't directly access the values by name until after instantiating, which again is not really that big of a deal, but I feel like I should avoid it.

This system works well enough, but it seems like it could be greatly improved.

Answer 1

You might think about using a database management system, like postgresql, or mysql. (And maybe even a fancy new noSql db ?) They worked and generally solved the problem of how to handle large data and still keep performance up...

If you devise your own format, you might end up at a point where you reimplement things like indexes (eg a btree) to improve performance.

That being said: For your simple example, I would say csv is suitable. Just add a header line (name;size;color) so that one can see the order of the columns from the header. This solves the hardcoding the order problem...

Answer 2

I've never used it, but check out pickle/cpickle . It might be easier to handle :

# Save a dictionary into a pickle file.
import pickle

favorite_color = { "lion": "yellow", "kitty": "red" }

pickle.dump( favorite_color, open( "save.p", "wb" ) )


#Load the dictionary back from the pickle file.

favorite_color = pickle.load( open( "save.p", "rb" ) )
# favorite_color is now { "lion": "yellow", "kitty": "red" }